Initial commit

.gitattributes

@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+jpg filter=lfs diff=lfs merge=lfs -text
+.jpg filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -0,0 +1,2 @@
+env
+*/__pycache__/*

LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright 2019 Sheng-Yu Wang, Oliver Wang, Andrew Owens, Richard Zhang, Alexei A. Efros
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

README.md

@@ -1,12 +1,82 @@
----
-title: P PD
-emoji: 📈
-colorFrom: green
-colorTo: indigo
-sdk: streamlit
-sdk_version: 1.25.0
-app_file: app.py
-pinned: false
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+## <b>Detecting Photoshopped Faces by Scripting Photoshop</b> <br>[[Project Page]](http://peterwang512.github.io/FALdetector) [[Paper]](https://arxiv.org/abs/1906.05856)
+
+[Sheng-Yu Wang<sup>1</sup>](https://peterwang512.github.io/),
+[Oliver Wang<sup>2</sup>](http://www.oliverwang.info/),
+[Andrew Owens<sup>1</sup>](http://andrewowens.com/),
+[Richard Zhang<sup>2</sup>](https://richzhang.github.io/),
+[Alexei A. Efros<sup>1</sup>](https://people.eecs.berkeley.edu/~efros/). <br>
+UC Berkeley<sup>1</sup>, Adobe Research<sup>2</sup>. <br>
+In [ICCV, 2019](https://arxiv.org/abs/1906.05856).
+
+
+<img src='https://peterwang512.github.io/FALdetector/images/teaser.png' align="center" width=900>
+
+<b>9/30/2019 Update</b> The code and model weights have been updated to correspond to the v2 of our paper. Note that the global classifer architecture is changed from resnet-50 to drn-c-26.
+
+<b>1/19/2019 Update</b> Dataset for evaluation is released! The link is [here](https://drive.google.com/file/d/1qCnwdbXFTf96g_LP-g_h-0BQcypNDuWB/view).
+
+## (0) Disclaimer
+Welcome! Computer vision algorithms often work well on some images, but fail on others. Ours is like this too. We believe our work is a significant step forward in detecting and undoing facial warping by image editing tools. However, there are still many hard cases, and this is by no means a solved problem.
+
+This is partly because our algorithm is trained on faces warped by the Face-aware Liquify tool in Photoshop, and will thus work well for these types of images, but not necessarily for others. We call this the "dataset bias" problem. Please see the paper for more details on this issue.
+
+While we trained our models with various data augmentation to be more robust to downstream operations such as resizing, jpeg compression and saturation/brightness changes, there are many other retouches (e.g. airbrushing) that can alter the low-level statistics of the images to make the detection a really hard one.
+
+Please enjoy our results and have fun trying out our models!
+
+
+
+
+## (1) Setup
+
+### Install packages
+- Install PyTorch ([pytorch.org](http://pytorch.org))
+- `pip install -r requirements.txt`
+
+### Download model weights
+- Run `bash weights/download_weights.sh`
+
+
+## (2) Run our models
+ 
+### Global classifer
+```
+python global_classifier.py --input_path examples/modified.jpg --model_path weights/global.pth
+```
+
+### Local Detector
+```
+python local_detector.py --input_path examples/modified.jpg --model_path weights/local.pth --dest_folder out/
+```
+
+**Note:** Our models are trained on faces cropped by the dlib CNN face detector. Although in both scripts we included the `--no_crop` option to run the models without face crops, it is used for images with already cropped faces.
+
+## (3) Dataset
+A validation set consisting of 500 original and 500 modified images each from Flickr and OpenImage can be downloaded [here](https://drive.google.com/file/d/1qCnwdbXFTf96g_LP-g_h-0BQcypNDuWB/view). Due to licensing issues, the released validation set is different from the set we evaluate in the paper, and the training set will not be released.
+
+In the zip file, original faces are in the `original` folder, and modified faces are in the `modified` folder. For reference, the `reference` folder contains the same faces in the `modified` folder, but those are before modification (original).
+
+To evaluate the dataset, run:
+```
+# Download the dataset
+cd data
+bash download_valset.sh
+cd ..
+# Run evaluation script. Model weights need to be downloaded.
+python eval.py --dataroot data --global_pth weights/global.pth --local_pth weights/local.pth --gpu_id 0
+```
+The following are the models' performances on the released set:
+
+|Accuracy|  AP |PSNR Increase|
+|:------:|:---:|:-----------:|
+|   93.9%|98.9%|        +2.66|
+
+
+
+## (A) Acknowledgments
+
+This repository borrows partially from the [pytorch-CycleGAN-and-pix2pix](https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix), [drn](https://github.com/fyu/drn), and the PyTorch [torchvision models](https://github.com/pytorch/vision/tree/master/torchvision/models) repositories. 
+
+## (B) Citation, Contact
+
+If you find this useful for your research, please consider citing this [bibtex](https://peterwang512.github.io/FALdetector/cite.txt). Please contact Sheng-Yu Wang \<sheng-yu_wang at berkeley dot edu\> with any comments or feedback.

app.py

@@ -0,0 +1,94 @@
+import base64
+import json
+import os, shutil
+import re
+import time
+import uuid
+
+import cv2
+
+import numpy as np
+import streamlit as st
+from PIL import Image
+# from extract_video import extract_method_single_video
+
+import shlex
+import subprocess
+from file_picker import st_file_selector
+
+import os
+
+from inference import classify_fake, heatmap_analysis
+
+DEBUG = True
+SAMPLE_FOLDER = 'examples'
+
+def main():
+    st.markdown("###")    
+    uploaded_file = st.file_uploader('Upload a picture', type=['jpg', 'jpeg', 'png'], accept_multiple_files=False)
+
+    with st.spinner(f'Loading samples...'):
+        while not os.path.isdir(SAMPLE_FOLDER):
+            time.sleep(1)
+    st.markdown("### or")
+    selected_file = st_file_selector(st, path=SAMPLE_FOLDER, key = 'selected', label = 'Choose a sample image')
+
+    if uploaded_file: 
+        img = Image.open(uploaded_file).convert('RGB')
+        st.image(img)
+    elif selected_file:
+        img = Image.open(os.path.join(SAMPLE_FOLDER, selected_file)).convert('RGB')
+        st.image(img)
+    else:
+        return
+
+   
+
+    
+
+    
+    
+    with st.spinner(f'Analyzing image...'):
+        try:
+            modified_probability = classify_fake(img)
+
+        except Exception as e:
+            if DEBUG:
+                st.write(e)
+            else:
+                st.text("Encountered a problem while analyzing image 🚨")
+                return
+            
+    if modified_probability > 0.6:
+        st.error('  MODIFIED IMAGE! ', icon="🚨")
+    else:
+        st.success("  REAL IMAGE! ", icon="✅")
+            
+    st.text("modified probability {:.2f}".format(modified_probability))
+
+
+    if modified_probability > 0.6:
+        with st.spinner(f'Analyzing heatmap...'):
+            try:
+                modified, reverse, heatmap = heatmap_analysis(img)
+
+            except Exception as e:
+                if DEBUG:
+                    st.write(e)
+                else:
+                    st.text("Encountered a problem while analyzing image 🚨")
+                    return
+        
+        st.write("### Heatmap")
+        st.image(heatmap)
+
+        st.write("### Reversed stretch imgae")
+        st.image(reverse)
+
+
+if __name__ == "__main__":
+    st.set_page_config(
+        page_title="Nodeflux Photosop Detection", page_icon=":pencil2:"
+    )
+    st.title("Photosop Detection")
+    main()

data/download_valset.sh

@@ -0,0 +1,9 @@
+#!/bin/bash
+DOCUMENT_ID="1mzNxCyrUTBF7-lQGPLYT0HuUODvVvtsb"
+FINAL_DOWNLOADED_FILENAME="val.zip"
+
+curl -c /tmp/cookies "https://drive.google.com/uc?export=download&id=$DOCUMENT_ID" > /tmp/intermezzo.html
+curl -L -b /tmp/cookies "https://drive.google.com$(cat /tmp/intermezzo.html | grep -Po 'uc-download-link" [^>]* href="\K[^"]*' | sed 's/\&/\&/g')" > $FINAL_DOWNLOADED_FILENAME
+unzip val.zip
+rm val.zip
+

eval.py

@@ -0,0 +1,119 @@
+import glob
+import argparse
+import torch
+import torchvision.transforms as transforms
+import numpy as np
+from PIL import Image
+
+from networks.drn_seg import DRNSeg, DRNSub
+from utils.tools import *
+from utils.visualize import *
+from sklearn.metrics import average_precision_score, accuracy_score
+
+
+def load_global_classifier(model_path, gpu_id):
+    if torch.cuda.is_available() and gpu_id != -1:
+        device = 'cuda:{}'.format(gpu_id)
+    else:
+        device = 'cpu'
+    model = DRNSub(1)
+    state_dict = torch.load(model_path, map_location='cpu')
+    model.load_state_dict(state_dict['model'])
+    model.to(device)
+    model.device = device
+    model.eval()
+    return model
+
+
+def load_local_detector(model_path, gpu_id):
+    if torch.cuda.is_available():
+        device = 'cuda:{}'.format(gpu_id)
+    else:
+        device = 'cpu'
+
+    model = DRNSeg(2)
+    state_dict = torch.load(model_path, map_location=device)
+    model.load_state_dict(state_dict['model'])
+    model.to(device)
+    model.device = device
+    model.eval()
+    return model
+
+
+tf = transforms.Compose([transforms.ToTensor(),
+                         transforms.Normalize(mean=[0.485, 0.456, 0.406],
+                                              std=[0.229, 0.224, 0.225])])
+def load_data(img_path, device):
+    face = Image.open(img_path).convert('RGB')
+    face = resize_shorter_side(face, 400)[0]
+    face_tens = tf(face).to(device)
+    return face_tens, face
+
+
+def classify_fake(model, img_path):
+    img = load_data(img_path, model.device)[0].unsqueeze(0)
+    # Prediction
+    with torch.no_grad():
+        prob = model(img)[0].sigmoid().cpu().item()
+    return prob
+
+
+def calc_psnr(img0, img1, mask=None):
+    return -10 * np.log10(np.mean((img0 - img1)**2) + 1e-6)
+
+
+def detect_warp(model, img_path):
+    img, modified = load_data(img_path, model.device)
+    # Warping field prediction
+    with torch.no_grad():
+        flow = model(img.unsqueeze(0))[0].cpu().numpy()
+        flow = np.transpose(flow, (1, 2, 0))
+
+    # Undoing the warps
+    flow = flow_resize(flow, modified.size)
+    modified_np = np.asarray(modified)
+    reverse_np = warp(modified_np, flow)
+    original = Image.open(img_path.replace('modified', 'reference')).convert('RGB')
+    original_np = np.asarray(original.resize(modified.size, Image.BICUBIC))
+
+    psnr_before = calc_psnr(original_np / 255, modified_np / 255)
+    psnr_after = calc_psnr(original_np / 255, reverse_np / 255)
+    return psnr_before, psnr_after
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser()
+    parser.add_argument(
+        "--dataroot", required=True, help='the root to the dataset')
+    parser.add_argument(
+        "--global_pth", required=True, help="path to the global model")
+    parser.add_argument(
+        "--local_pth", required=True, help="path to the local model")
+    parser.add_argument(
+        "--gpu_id", default='0', help="the id of the gpu to run model on")
+    args = parser.parse_args()
+
+    glb_model = load_global_classifier(args.global_pth, args.gpu_id)
+    lcl_model = load_local_detector(args.local_pth, args.gpu_id)
+
+    pred_prob, gt_prob, psnr_before, psnr_after = [], [], [], []
+    for img_path in glob.glob(args.dataroot + '/original/*'):
+        pred_prob.append(classify_fake(glb_model, img_path))
+        gt_prob.append(0)
+
+    for img_path in glob.glob(args.dataroot + '/modified/*'):
+        pred_prob.append(classify_fake(glb_model, img_path))
+        gt_prob.append(1)
+        psnrs = detect_warp(lcl_model, img_path)
+        psnr_before.append(psnrs[0])
+        psnr_after.append(psnrs[1])
+
+    pred_prob, gt_prob, psnr_before, psnr_after = \
+        np.array(pred_prob), np.array(gt_prob), np.array(psnr_before), np.array(psnr_after)
+    acc = accuracy_score(gt_prob, pred_prob > 0.5)
+    avg_precision = average_precision_score(gt_prob, pred_prob)
+    delta_psnr = psnr_after.mean() - psnr_before.mean()
+
+    print("Accuracy: ", acc)
+    print("Average precision: ", avg_precision)
+    print("PSNR increase: ", delta_psnr)

file_picker.py

@@ -0,0 +1,47 @@
+"""FilePicker for streamlit. 
+Still doesn't seem to be a good solution for a way to select files to process from the server Streamlit is running on.
+Here's a pretty functional solution. 
+Usage:
+```
+import streamlit as st
+from filepicker import st_file_selector
+tif_file = st_file_selector(st, key = 'tif', label = 'Choose tif file')
+```
+"""
+
+import os
+import streamlit as st
+
+def update_dir(key):
+    choice = st.session_state[key]
+    if os.path.isdir(os.path.join(st.session_state[key+'curr_dir'], choice)):
+        st.session_state[key+'curr_dir'] = os.path.normpath(os.path.join(st.session_state[key+'curr_dir'], choice))
+        files = sorted(os.listdir(st.session_state[key+'curr_dir']))
+        if "images" in files:
+          files.remove("images")
+        st.session_state[key+'files'] = files
+
+def st_file_selector(st_placeholder, path='.', label='Select a file/folder', key = 'selected'):
+    if key+'curr_dir' not in st.session_state:
+        base_path = '.' if path is None or path == '' else path
+        base_path = base_path if os.path.isdir(base_path) else os.path.dirname(base_path)
+        base_path = '.' if base_path is None or base_path == '' else base_path
+
+        files = sorted(os.listdir(base_path))
+        files.insert(0, 'Choose a file...')
+        if "images" in files:
+          files.remove("images")
+        st.session_state[key+'files'] = files
+        st.session_state[key+'curr_dir'] = base_path
+    else:
+        base_path = st.session_state[key+'curr_dir']
+
+    selected_file = st_placeholder.selectbox(label=label, 
+                                        options=st.session_state[key+'files'], 
+                                        key=key, 
+                                        on_change = lambda: update_dir(key))
+    
+    if selected_file == "Choose a file...":
+        return None
+
+    return selected_file

global_classifier.py

@@ -0,0 +1,67 @@
+import argparse
+import os
+import sys
+import torch
+from PIL import Image
+import torchvision.transforms as transforms
+
+from networks.drn_seg import DRNSub
+from utils.tools import *
+from utils.visualize import *
+
+
+def load_classifier(model_path, gpu_id):
+    if torch.cuda.is_available() and gpu_id != -1:
+        device = 'cuda:{}'.format(gpu_id)
+    else:
+        device = 'cpu'
+    model = DRNSub(1)
+    state_dict = torch.load(model_path, map_location='cpu')
+    model.load_state_dict(state_dict['model'])
+    model.to(device)
+    model.device = device
+    model.eval()
+    return model
+
+
+tf = transforms.Compose([transforms.ToTensor(),
+                         transforms.Normalize(mean=[0.485, 0.456, 0.406],
+                                              std=[0.229, 0.224, 0.225])])
+def classify_fake(model, img_path, no_crop=False,
+                  model_file='utils/dlib_face_detector/mmod_human_face_detector.dat'):
+    # Data preprocessing
+    im_w, im_h = Image.open(img_path).size
+    if no_crop:
+        face = Image.open(img_path).convert('RGB')
+    else:
+        faces = face_detection(img_path, verbose=False, model_file=model_file)
+        if len(faces) == 0:
+            print("no face detected by dlib, exiting")
+            sys.exit()
+        face, box = faces[0]
+    face = resize_shorter_side(face, 400)[0]
+    face_tens = tf(face).to(model.device)
+
+    # Prediction
+    with torch.no_grad():
+        prob = model(face_tens.unsqueeze(0))[0].sigmoid().cpu().item()
+    return prob
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser()
+    parser.add_argument(
+        "--input_path", required=True, help="the model input")
+    parser.add_argument(
+        "--model_path", required=True, help="path to the drn model")
+    parser.add_argument(
+        "--gpu_id", default='0', help="the id of the gpu to run model on")
+    parser.add_argument(
+        "--no_crop",
+        action="store_true",
+        help="do not use a face detector, instead run on the full input image")
+    args = parser.parse_args()
+
+    model = load_classifier(args.model_path, args.gpu_id)
+    prob = classify_fake(model, args.input_path, args.no_crop)
+    print("Probibility being modified by Photoshop FAL: {:.2f}%".format(prob*100))

inference.py

@@ -0,0 +1,119 @@
+import argparse
+import os
+import sys
+import numpy as np
+import torch
+import torchvision.transforms as transforms
+from PIL import Image
+
+from networks.drn_seg import DRNSeg, DRNSub
+from utils.tools import *
+from utils.visualize import *
+
+def load_classifier(model_path, gpu_id):
+    if torch.cuda.is_available() and gpu_id != -1:
+        device = 'cuda:{}'.format(gpu_id)
+    else:
+        device = 'cpu'
+    model = DRNSub(1)
+    state_dict = torch.load(model_path, map_location='cpu')
+    model.load_state_dict(state_dict['model'])
+    model.to(device)
+    model.device = device
+    model.eval()
+    return model
+
+
+
+local_model_path = 'weights/local.pth'
+global_model_path = 'weights/global.pth'
+gpu_id = 0
+
+# Loading the model
+if torch.cuda.is_available():
+    device = 'cuda:{}'.format(gpu_id)
+else:
+    device = 'cpu'
+
+local_model = DRNSeg(2)
+state_dict = torch.load(local_model_path, map_location=device)
+local_model.load_state_dict(state_dict['model'])
+local_model.to(device)
+local_model.eval()
+
+global_model = load_classifier(global_model_path, gpu_id)
+
+# prob = classify_fake(model, args.input_path, args.no_crop)
+
+# Data preprocessing
+tf = transforms.Compose([
+    transforms.ToTensor(),
+    transforms.Normalize(
+        mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
+])
+
+
+def classify_fake(img, no_crop=False, global_model=global_model,
+                  model_file='utils/dlib_face_detector/mmod_human_face_detector.dat'):
+    # Data preprocessing
+    im_w, im_h = img.size
+    if no_crop:
+        face = img
+    else:
+        faces = face_detection(img, verbose=False, model_file=model_file)
+        if len(faces) == 0:
+            print("no face detected by dlib, exiting")
+            sys.exit()
+        face, box = faces[0]
+    face = resize_shorter_side(face, 400)[0]
+    face_tens = tf(face).to(global_model.device)
+
+    # Prediction
+    with torch.no_grad():
+        prob = global_model(face_tens.unsqueeze(0))[0].sigmoid().cpu().item()
+    return prob
+
+
+
+def heatmap_analysis(img, no_crop=False):
+
+    im_w, im_h = img.size
+    if no_crop:
+        face = imgs
+    else:
+        faces = face_detection(img, verbose=False)
+        if len(faces) == 0:
+            print("no face detected by dlib, exiting")
+            sys.exit()
+        face, box = faces[0]
+    face = resize_shorter_side(face, 400)[0]
+    face_tens = tf(face).to(device)
+
+    # Warping field prediction
+    with torch.no_grad():
+        flow = local_model(face_tens.unsqueeze(0))[0].cpu().numpy()
+        flow = np.transpose(flow, (1, 2, 0))
+        h, w, _ = flow.shape
+
+    # Undoing the warps
+    modified = face.resize((w, h), Image.BICUBIC)
+    modified_np = np.asarray(modified)
+    reverse_np = warp(modified_np, flow)
+    reverse = Image.fromarray(reverse_np)
+
+    flow_magn = np.sqrt(flow[:, :, 0]**2 + flow[:, :, 1]**2)
+    cv_out = get_heatmap_cv(modified_np, flow_magn, 7)
+    heatmap = Image.fromarray(cv_out)
+    return modified, reverse, heatmap
+
+    # Saving the results
+    # modified.save(
+    #     os.path.join(dest_folder, 'cropped_input.jpg'),
+    #     quality=90)
+    # reverse.save(
+    #     os.path.join(dest_folder, 'warped.jpg'),
+    #     quality=90)
+    # flow_magn = np.sqrt(flow[:, :, 0]**2 + flow[:, :, 1]**2)
+    # save_heatmap_cv(
+    #     modified_np, flow_magn,
+    #     os.path.join(dest_folder, 'heatmap.jpg'))

local_detector.py

@@ -0,0 +1,88 @@
+import argparse
+import os
+import sys
+import numpy as np
+import torch
+import torchvision.transforms as transforms
+from PIL import Image
+
+from networks.drn_seg import DRNSeg
+from utils.tools import *
+from utils.visualize import *
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser()
+    parser.add_argument(
+        "--input_path", required=True, help="the model input")
+    parser.add_argument(
+        "--dest_folder", required=True, help="folder to store the results")
+    parser.add_argument(
+        "--model_path", required=True, help="path to the drn model")
+    parser.add_argument(
+        "--gpu_id", default='0', help="the id of the gpu to run model on")
+    parser.add_argument(
+        "--no_crop",
+        action="store_true",
+        help="do not use a face detector, instead run on the full input image")
+    args = parser.parse_args()
+
+    img_path = args.input_path
+    dest_folder = args.dest_folder
+    model_path = args.model_path
+    gpu_id = args.gpu_id
+
+    # Loading the model
+    if torch.cuda.is_available():
+        device = 'cuda:{}'.format(gpu_id)
+    else:
+        device = 'cpu'
+
+    model = DRNSeg(2)
+    state_dict = torch.load(model_path, map_location=device)
+    model.load_state_dict(state_dict['model'])
+    model.to(device)
+    model.eval()
+
+    # Data preprocessing
+    tf = transforms.Compose([
+        transforms.ToTensor(),
+        transforms.Normalize(
+            mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
+    ])
+
+    im_w, im_h = Image.open(img_path).size
+    if args.no_crop:
+        face = Image.open(img_path).convert('RGB')
+    else:
+        faces = face_detection(img_path, verbose=False)
+        if len(faces) == 0:
+            print("no face detected by dlib, exiting")
+            sys.exit()
+        face, box = faces[0]
+    face = resize_shorter_side(face, 400)[0]
+    face_tens = tf(face).to(device)
+
+    # Warping field prediction
+    with torch.no_grad():
+        flow = model(face_tens.unsqueeze(0))[0].cpu().numpy()
+        flow = np.transpose(flow, (1, 2, 0))
+        h, w, _ = flow.shape
+
+    # Undoing the warps
+    modified = face.resize((w, h), Image.BICUBIC)
+    modified_np = np.asarray(modified)
+    reverse_np = warp(modified_np, flow)
+    reverse = Image.fromarray(reverse_np)
+
+    # Saving the results
+    modified.save(
+        os.path.join(dest_folder, 'cropped_input.jpg'),
+        quality=90)
+    reverse.save(
+        os.path.join(dest_folder, 'warped.jpg'),
+        quality=90)
+    flow_magn = np.sqrt(flow[:, :, 0]**2 + flow[:, :, 1]**2)
+    save_heatmap_cv(
+        modified_np, flow_magn,
+        os.path.join(dest_folder, 'heatmap.jpg'))

networks/drn.py

@@ -0,0 +1,416 @@
+import pdb
+
+import torch
+import torch.nn as nn
+import math
+import torch.utils.model_zoo as model_zoo
+
+torch.backends.cudnn.benchmark = True
+BatchNorm = nn.BatchNorm2d
+
+
+# __all__ = ['DRN', 'drn26', 'drn42', 'drn58']
+
+
+webroot = 'https://tigress-web.princeton.edu/~fy/drn/models/'
+
+model_urls = {
+    'resnet50': 'https://download.pytorch.org/models/resnet50-19c8e357.pth',
+    'drn-c-26': webroot + 'drn_c_26-ddedf421.pth',
+    'drn-c-42': webroot + 'drn_c_42-9d336e8c.pth',
+    'drn-c-58': webroot + 'drn_c_58-0a53a92c.pth',
+    'drn-d-22': webroot + 'drn_d_22-4bd2f8ea.pth',
+    'drn-d-38': webroot + 'drn_d_38-eebb45f0.pth',
+    'drn-d-54': webroot + 'drn_d_54-0e0534ff.pth',
+    'drn-d-105': webroot + 'drn_d_105-12b40979.pth'
+}
+
+
+def conv3x3(in_planes, out_planes, stride=1, padding=1, dilation=1):
+    return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride,
+                     padding=padding, bias=False, dilation=dilation)
+
+
+class BasicBlock(nn.Module):
+    expansion = 1
+
+    def __init__(self, inplanes, planes, stride=1, downsample=None,
+                 dilation=(1, 1), residual=True):
+        super(BasicBlock, self).__init__()
+        self.conv1 = conv3x3(inplanes, planes, stride,
+                             padding=dilation[0], dilation=dilation[0])
+        self.bn1 = BatchNorm(planes)
+        self.relu = nn.ReLU(inplace=True)
+        self.conv2 = conv3x3(planes, planes,
+                             padding=dilation[1], dilation=dilation[1])
+        self.bn2 = BatchNorm(planes)
+        self.downsample = downsample
+        self.stride = stride
+        self.residual = residual
+
+    def forward(self, x):
+        residual = x
+
+        out = self.conv1(x)
+        out = self.bn1(out)
+        out = self.relu(out)
+
+        out = self.conv2(out)
+        out = self.bn2(out)
+
+        if self.downsample is not None:
+            residual = self.downsample(x)
+        if self.residual:
+            out += residual
+        out = self.relu(out)
+
+        return out
+
+
+class Bottleneck(nn.Module):
+    expansion = 4
+
+    def __init__(self, inplanes, planes, stride=1, downsample=None,
+                 dilation=(1, 1), residual=True):
+        super(Bottleneck, self).__init__()
+        self.conv1 = nn.Conv2d(inplanes, planes, kernel_size=1, bias=False)
+        self.bn1 = BatchNorm(planes)
+        self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=stride,
+                               padding=dilation[1], bias=False,
+                               dilation=dilation[1])
+        self.bn2 = BatchNorm(planes)
+        self.conv3 = nn.Conv2d(planes, planes * 4, kernel_size=1, bias=False)
+        self.bn3 = BatchNorm(planes * 4)
+        self.relu = nn.ReLU(inplace=True)
+        self.downsample = downsample
+        self.stride = stride
+
+    def forward(self, x):
+        residual = x
+
+        out = self.conv1(x)
+        out = self.bn1(out)
+        out = self.relu(out)
+
+        out = self.conv2(out)
+        out = self.bn2(out)
+        out = self.relu(out)
+
+        out = self.conv3(out)
+        out = self.bn3(out)
+
+        if self.downsample is not None:
+            residual = self.downsample(x)
+
+        out += residual
+        out = self.relu(out)
+
+        return out
+
+
+class DRN(nn.Module):
+
+    def __init__(self, block, layers, num_classes=1000,
+                 channels=(16, 32, 64, 128, 256, 512, 512, 512),
+                 out_map=False, out_middle=False, pool_size=28, arch='D'):
+        super(DRN, self).__init__()
+        self.inplanes = channels[0]
+        self.out_map = out_map
+        self.out_dim = channels[-1]
+        self.out_middle = out_middle
+        self.arch = arch
+
+        if arch == 'C':
+            self.conv1 = nn.Conv2d(3, channels[0], kernel_size=7, stride=1,
+                                   padding=3, bias=False)
+            self.bn1 = BatchNorm(channels[0])
+            self.relu = nn.ReLU(inplace=True)
+
+            self.layer1 = self._make_layer(
+                BasicBlock, channels[0], layers[0], stride=1)
+            self.layer2 = self._make_layer(
+                BasicBlock, channels[1], layers[1], stride=2)
+        elif arch == 'D':
+            self.layer0 = nn.Sequential(
+                nn.Conv2d(3, channels[0], kernel_size=7, stride=1, padding=3,
+                          bias=False),
+                BatchNorm(channels[0]),
+                nn.ReLU(inplace=True)
+            )
+
+            self.layer1 = self._make_conv_layers(
+                channels[0], layers[0], stride=1)
+            self.layer2 = self._make_conv_layers(
+                channels[1], layers[1], stride=2)
+
+        self.layer3 = self._make_layer(block, channels[2], layers[2], stride=2)
+        self.layer4 = self._make_layer(block, channels[3], layers[3], stride=2)
+        self.layer5 = self._make_layer(block, channels[4], layers[4],
+                                       dilation=2, new_level=False)
+        self.layer6 = None if layers[5] == 0 else \
+            self._make_layer(block, channels[5], layers[5], dilation=4,
+                             new_level=False)
+
+        if arch == 'C':
+            self.layer7 = None if layers[6] == 0 else \
+                self._make_layer(BasicBlock, channels[6], layers[6], dilation=2,
+                                 new_level=False, residual=False)
+            self.layer8 = None if layers[7] == 0 else \
+                self._make_layer(BasicBlock, channels[7], layers[7], dilation=1,
+                                 new_level=False, residual=False)
+        elif arch == 'D':
+            self.layer7 = None if layers[6] == 0 else \
+                self._make_conv_layers(channels[6], layers[6], dilation=2)
+            self.layer8 = None if layers[7] == 0 else \
+                self._make_conv_layers(channels[7], layers[7], dilation=1)
+
+        if num_classes > 0:
+            self.avgpool = nn.AvgPool2d(pool_size)
+            self.fc = nn.Conv2d(self.out_dim, num_classes, kernel_size=1,
+                                stride=1, padding=0, bias=True)
+        for m in self.modules():
+            if isinstance(m, nn.Conv2d):
+                n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
+                m.weight.data.normal_(0, math.sqrt(2. / n))
+            elif isinstance(m, BatchNorm):
+                m.weight.data.fill_(1)
+                m.bias.data.zero_()
+
+    def _make_layer(self, block, planes, blocks, stride=1, dilation=1,
+                    new_level=True, residual=True):
+        assert dilation == 1 or dilation % 2 == 0
+        downsample = None
+        if stride != 1 or self.inplanes != planes * block.expansion:
+            downsample = nn.Sequential(
+                nn.Conv2d(self.inplanes, planes * block.expansion,
+                          kernel_size=1, stride=stride, bias=False),
+                BatchNorm(planes * block.expansion),
+            )
+
+        layers = list()
+        layers.append(block(
+            self.inplanes, planes, stride, downsample,
+            dilation=(1, 1) if dilation == 1 else (
+                dilation // 2 if new_level else dilation, dilation),
+            residual=residual))
+        self.inplanes = planes * block.expansion
+        for i in range(1, blocks):
+            layers.append(block(self.inplanes, planes, residual=residual,
+                                dilation=(dilation, dilation)))
+
+        return nn.Sequential(*layers)
+
+    def _make_conv_layers(self, channels, convs, stride=1, dilation=1):
+        modules = []
+        for i in range(convs):
+            modules.extend([
+                nn.Conv2d(self.inplanes, channels, kernel_size=3,
+                          stride=stride if i == 0 else 1,
+                          padding=dilation, bias=False, dilation=dilation),
+                BatchNorm(channels),
+                nn.ReLU(inplace=True)])
+            self.inplanes = channels
+        return nn.Sequential(*modules)
+
+    def forward(self, x):
+        y = list()
+
+        if self.arch == 'C':
+            x = self.conv1(x)
+            x = self.bn1(x)
+            x = self.relu(x)
+        elif self.arch == 'D':
+            x = self.layer0(x)
+
+        x = self.layer1(x)
+        y.append(x)
+        x = self.layer2(x)
+        y.append(x)
+
+        x = self.layer3(x)
+        y.append(x)
+
+        x = self.layer4(x)
+        y.append(x)
+
+        x = self.layer5(x)
+        y.append(x)
+
+        if self.layer6 is not None:
+            x = self.layer6(x)
+            y.append(x)
+
+        if self.layer7 is not None:
+            x = self.layer7(x)
+            y.append(x)
+
+        if self.layer8 is not None:
+            x = self.layer8(x)
+            y.append(x)
+
+        if self.out_map:
+            x = self.fc(x)
+        else:
+            x = self.avgpool(x)
+            x = self.fc(x)
+            x = x.view(x.size(0), -1)
+
+        if self.out_middle:
+            return x, y
+        else:
+            return x
+
+
+class DRN_A(nn.Module):
+
+    def __init__(self, block, layers, num_classes=1000):
+        self.inplanes = 64
+        super(DRN_A, self).__init__()
+        self.out_dim = 512 * block.expansion
+        self.conv1 = nn.Conv2d(3, 64, kernel_size=7, stride=2, padding=3,
+                               bias=False)
+        self.bn1 = nn.BatchNorm2d(64)
+        self.relu = nn.ReLU(inplace=True)
+        self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
+        self.layer1 = self._make_layer(block, 64, layers[0])
+        self.layer2 = self._make_layer(block, 128, layers[1], stride=2)
+        self.layer3 = self._make_layer(block, 256, layers[2], stride=1,
+                                       dilation=2)
+        self.layer4 = self._make_layer(block, 512, layers[3], stride=1,
+                                       dilation=4)
+        self.avgpool = nn.AvgPool2d(28, stride=1)
+        self.fc = nn.Linear(512 * block.expansion, num_classes)
+
+        for m in self.modules():
+            if isinstance(m, nn.Conv2d):
+                n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
+                m.weight.data.normal_(0, math.sqrt(2. / n))
+            elif isinstance(m, BatchNorm):
+                m.weight.data.fill_(1)
+                m.bias.data.zero_()
+
+        # for m in self.modules():
+        #     if isinstance(m, nn.Conv2d):
+        #         nn.init.kaiming_normal_(m.weight, mode='fan_out', nonlinearity='relu')
+        #     elif isinstance(m, nn.BatchNorm2d):
+        #         nn.init.constant_(m.weight, 1)
+        #         nn.init.constant_(m.bias, 0)
+
+    def _make_layer(self, block, planes, blocks, stride=1, dilation=1):
+        downsample = None
+        if stride != 1 or self.inplanes != planes * block.expansion:
+            downsample = nn.Sequential(
+                nn.Conv2d(self.inplanes, planes * block.expansion,
+                          kernel_size=1, stride=stride, bias=False),
+                nn.BatchNorm2d(planes * block.expansion),
+            )
+
+        layers = []
+        layers.append(block(self.inplanes, planes, stride, downsample))
+        self.inplanes = planes * block.expansion
+        for i in range(1, blocks):
+            layers.append(block(self.inplanes, planes,
+                                dilation=(dilation, dilation)))
+
+        return nn.Sequential(*layers)
+
+    def forward(self, x):
+        x = self.conv1(x)
+        x = self.bn1(x)
+        x = self.relu(x)
+        x = self.maxpool(x)
+
+        x = self.layer1(x)
+        x = self.layer2(x)
+        x = self.layer3(x)
+        x = self.layer4(x)
+
+        x = self.avgpool(x)
+        x = x.view(x.size(0), -1)
+        x = self.fc(x)
+
+        return x
+
+
+def drn_a_50(pretrained=False, **kwargs):
+    model = DRN_A(Bottleneck, [3, 4, 6, 3], **kwargs)
+    if pretrained:
+        model.load_state_dict(model_zoo.load_url(model_urls['resnet50']))
+    return model
+
+
+def drn_c_26(pretrained=False, **kwargs):
+    model = DRN(BasicBlock, [1, 1, 2, 2, 2, 2, 1, 1], arch='C', **kwargs)
+    if pretrained:
+        model.load_state_dict(model_zoo.load_url(model_urls['drn-c-26']))
+    return model
+
+
+def drn_c_42(pretrained=False, **kwargs):
+    model = DRN(BasicBlock, [1, 1, 3, 4, 6, 3, 1, 1], arch='C', **kwargs)
+    if pretrained:
+        model.load_state_dict(model_zoo.load_url(model_urls['drn-c-42']))
+    return model
+
+
+def drn_c_58(pretrained=False, **kwargs):
+    model = DRN(Bottleneck, [1, 1, 3, 4, 6, 3, 1, 1], arch='C', **kwargs)
+    if pretrained:
+        model.load_state_dict(model_zoo.load_url(model_urls['drn-c-58']))
+    return model
+
+
+def drn_d_22(pretrained=False, **kwargs):
+    model = DRN(BasicBlock, [1, 1, 2, 2, 2, 2, 1, 1], arch='D', **kwargs)
+    if pretrained:
+        model.load_state_dict(model_zoo.load_url(model_urls['drn-d-22']))
+    return model
+
+
+def drn_d_24(pretrained=False, **kwargs):
+    model = DRN(BasicBlock, [1, 1, 2, 2, 2, 2, 2, 2], arch='D', **kwargs)
+    if pretrained:
+        model.load_state_dict(model_zoo.load_url(model_urls['drn-d-24']))
+    return model
+
+
+def drn_d_38(pretrained=False, **kwargs):
+    model = DRN(BasicBlock, [1, 1, 3, 4, 6, 3, 1, 1], arch='D', **kwargs)
+    if pretrained:
+        model.load_state_dict(model_zoo.load_url(model_urls['drn-d-38']))
+    return model
+
+
+def drn_d_40(pretrained=False, **kwargs):
+    model = DRN(BasicBlock, [1, 1, 3, 4, 6, 3, 2, 2], arch='D', **kwargs)
+    if pretrained:
+        model.load_state_dict(model_zoo.load_url(model_urls['drn-d-40']))
+    return model
+
+
+def drn_d_54(pretrained=False, **kwargs):
+    model = DRN(Bottleneck, [1, 1, 3, 4, 6, 3, 1, 1], arch='D', **kwargs)
+    if pretrained:
+        model.load_state_dict(model_zoo.load_url(model_urls['drn-d-54']))
+    return model
+
+
+def drn_d_56(pretrained=False, **kwargs):
+    model = DRN(Bottleneck, [1, 1, 3, 4, 6, 3, 2, 2], arch='D', **kwargs)
+    if pretrained:
+        model.load_state_dict(model_zoo.load_url(model_urls['drn-d-56']))
+    return model
+
+
+def drn_d_105(pretrained=False, **kwargs):
+    model = DRN(Bottleneck, [1, 1, 3, 4, 23, 3, 1, 1], arch='D', **kwargs)
+    if pretrained:
+        model.load_state_dict(model_zoo.load_url(model_urls['drn-d-105']))
+    return model
+
+
+def drn_d_107(pretrained=False, **kwargs):
+    model = DRN(Bottleneck, [1, 1, 3, 4, 23, 3, 2, 2], arch='D', **kwargs)
+    if pretrained:
+        model.load_state_dict(model_zoo.load_url(model_urls['drn-d-107']))
+    return model

networks/drn_seg.py

@@ -0,0 +1,95 @@
+import math
+import torch
+import torch.nn as nn
+from networks.drn import drn_c_26
+
+
+def fill_up_weights(up):
+    w = up.weight.data
+    f = math.ceil(w.size(2) / 2)
+    c = (2 * f - 1 - f % 2) / (2. * f)
+    for i in range(w.size(2)):
+        for j in range(w.size(3)):
+            w[0, 0, i, j] = \
+                (1 - math.fabs(i / f - c)) * (1 - math.fabs(j / f - c))
+    for c in range(1, w.size(0)):
+        w[c, 0, :, :] = w[0, 0, :, :]
+
+
+class DRNSeg(nn.Module):
+    def __init__(self, classes, pretrained_drn=False,
+            pretrained_model=None, use_torch_up=False):
+        super(DRNSeg, self).__init__()
+
+        model = drn_c_26(pretrained=pretrained_drn)
+        self.base = nn.Sequential(*list(model.children())[:-2])
+        if pretrained_model:
+            self.load_pretrained(pretrained_model)
+
+        self.seg = nn.Conv2d(model.out_dim, classes,
+                             kernel_size=1, bias=True)
+
+        m = self.seg
+        n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
+        m.weight.data.normal_(0, math.sqrt(2. / n))
+        m.bias.data.zero_()
+        if use_torch_up:
+            self.up = nn.UpsamplingBilinear2d(scale_factor=8)
+        else:
+            up = nn.ConvTranspose2d(classes, classes, 16, stride=8, padding=4,
+                                    output_padding=0, groups=classes,
+                                    bias=False)
+            fill_up_weights(up)
+            up.weight.requires_grad = False
+            self.up = up
+
+    def forward(self, x):
+        x = self.base(x)
+        x = self.seg(x)
+        y = self.up(x)
+        return y
+
+    def optim_parameters(self, memo=None):
+        for param in self.base.parameters():
+            yield param
+        for param in self.seg.parameters():
+            yield param
+
+    def load_pretrained(self, pretrained_model):
+        print("loading the pretrained drn model from %s" % pretrained_model)
+        state_dict = torch.load(pretrained_model, map_location='cpu')
+        if hasattr(state_dict, '_metadata'):
+            del state_dict._metadata
+
+        # filter out unnecessary keys
+        pretrained_dict = state_dict['model']
+        pretrained_dict = {k[5:]: v for k, v in pretrained_dict.items() if k.split('.')[0] == 'base'}
+
+        # load the pretrained state dict
+        self.base.load_state_dict(pretrained_dict)
+
+
+class DRNSub(nn.Module):
+    def __init__(self, num_classes, pretrained_model=None, fix_base=False):
+        super(DRNSub, self).__init__()
+
+        drnseg = DRNSeg(2)
+        if pretrained_model:
+            print("loading the pretrained drn model from %s" % pretrained_model)
+            state_dict = torch.load(pretrained_model, map_location='cpu')
+            drnseg.load_state_dict(state_dict['model'])
+
+        self.base = drnseg.base
+        if fix_base:
+            for param in self.base.parameters():
+                param.requires_grad = False
+
+        self.avgpool = nn.AdaptiveAvgPool2d((1, 1))
+        self.fc = nn.Linear(512, num_classes)
+
+    def forward(self, x):
+        x = self.base(x)
+        x = self.avgpool(x)
+        x = x.view(x.size(0), -1)
+        x = self.fc(x)
+        return x

requirements.txt

@@ -0,0 +1,10 @@
+streamlit
+dlib
+mmcv
+scipy
+numpy
+matplotlib
+opencv_python
+Pillow
+torch>=0.4.0
+torchvision

utils/tools.py

@@ -0,0 +1,140 @@
+import os
+import cv2
+import torch
+import numpy as np
+from PIL import Image
+from dlib import cnn_face_detection_model_v1 as face_detect_model
+
+
+def center_crop(im, length):
+    w, h = im.size
+    left = w//2 - length//2
+    right = w//2 + length//2
+    top = h//2 - length//2
+    bottom = h//2 + length//2
+    return im.crop((left, top, right, bottom)), (left, top)
+
+
+def remove_boundary(img):
+    """
+    Remove boundary artifacts that FAL causes.
+    """
+    w, h = img.size
+    left = w//80
+    top = h//50
+    right = w*79//80
+    bottom = h*24//25
+    return img.crop((left, top, right, bottom))
+
+
+def resize_shorter_side(img, min_length):
+    """
+    Resize the shorter side of img to min_length while
+    preserving the aspect ratio.
+    """
+    ow, oh = img.size
+    mult = 8
+    if ow < oh:
+        if ow == min_length and oh % mult == 0:
+            return img, (ow, oh)
+        w = min_length
+        h = int(min_length * oh / ow)
+    else:
+        if oh == min_length and ow % mult == 0:
+            return img, (ow, oh)
+        h = min_length
+        w = int(min_length * ow / oh)
+    return img.resize((w, h), Image.BICUBIC), (w, h)
+
+
+def flow_resize(flow, sz):
+    oh, ow, _ = flow.shape
+    w, h = sz
+    u_ = cv2.resize(flow[:,:,0], (w, h))
+    v_ = cv2.resize(flow[:,:,1], (w, h))
+    u_ *= w / float(ow)
+    v_ *= h / float(oh)
+    return np.dstack((u_,v_))
+
+
+def warp(im, flow, alpha=1, interp=cv2.INTER_CUBIC):
+    height, width, _ = flow.shape
+    cart = np.dstack(np.meshgrid(np.arange(width), np.arange(height)))
+    pixel_map = (cart + alpha * flow).astype(np.float32)
+    warped = cv2.remap(
+        im,
+        pixel_map[:, :, 0],
+        pixel_map[:, :, 1],
+        interp,
+        borderMode=cv2.BORDER_REPLICATE)
+    return warped
+
+
+cnn_face_detector = None
+def face_detection(
+        img,
+        verbose=False,
+        model_file='utils/dlib_face_detector/mmod_human_face_detector.dat'):
+    """
+    Detects faces using dlib cnn face detection, and extend the bounding box
+    to include the entire face.
+    """
+    def shrink(img, max_length=2048):
+        ow, oh = img.size
+        if max_length >= max(ow, oh):
+            return img, 1.0
+
+        if ow > oh:
+            mult = max_length / ow
+        else:
+            mult = max_length / oh
+        w = int(ow * mult)
+        h = int(oh * mult)
+        return img.resize((w, h), Image.BILINEAR), mult
+
+    global cnn_face_detector
+    if cnn_face_detector is None:
+        cnn_face_detector = face_detect_model(model_file)
+
+    w, h = img.size
+    img_shrinked, mult = shrink(img)
+
+    im = np.asarray(img_shrinked)
+    if len(im.shape) != 3 or im.shape[2] != 3:
+        return []
+
+    crop_ims = []
+    dets = cnn_face_detector(im, 0)
+    for k, d in enumerate(dets):
+        top = d.rect.top() / mult
+        bottom = d.rect.bottom() / mult
+        left = d.rect.left() / mult
+        right = d.rect.right() / mult
+
+        wid = right - left
+        left = max(0, left - wid // 2.5)
+        top = max(0, top - wid // 1.5)
+        right = min(w - 1, right + wid // 2.5)
+        bottom = min(h - 1, bottom + wid // 2.5)
+
+        if d.confidence > 1:
+            if verbose:
+                print("%d-th face detected: (%d, %d, %d, %d)" %
+                      (k, left, top, right, bottom))
+            crop_im = img.crop((left, top, right, bottom))
+            crop_ims.append((crop_im, (left, top, right, bottom)))
+
+    return crop_ims
+
+
+def mkdirs(paths):
+    if isinstance(paths, list) and not isinstance(paths, str):
+        for path in paths:
+            mkdir(path)
+    else:
+        mkdir(paths)
+
+
+def mkdir(path):
+    if not os.path.exists(path):
+        os.makedirs(path)

utils/visualize.py

@@ -0,0 +1,59 @@
+import os
+import cv2
+import torch
+import numpy as np
+import torchvision
+from PIL import Image
+
+
+def unnormalize(tens, mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]):
+    # assume tensor of shape NxCxHxW
+    return tens * torch.Tensor(std)[None, :, None, None] + torch.Tensor(
+        mean)[None, :, None, None]
+
+
+def get_heatmap_cv(img, magn, max_flow_mag):
+    min_flow_mag = .5
+    cv_magn = np.clip(
+        255 * (magn - min_flow_mag) / (max_flow_mag - min_flow_mag),
+        a_min=0,
+        a_max=255).astype(np.uint8)
+    if img.dtype != np.uint8:
+        img = (255 * img).astype(np.uint8)
+
+    heatmap_img = cv2.applyColorMap(cv_magn, cv2.COLORMAP_JET)
+    heatmap_img = heatmap_img[..., ::-1]
+
+    h, w = magn.shape
+    img_alpha = np.ones((h, w), dtype=np.double)[:, :, None]
+    heatmap_alpha = np.clip(
+        magn / max_flow_mag, a_min=0, a_max=1)[:, :, None]**.7
+    heatmap_alpha[heatmap_alpha < .2]**.5
+    pm_hm = heatmap_img * heatmap_alpha
+    pm_img = img * img_alpha
+    cv_out = pm_hm + pm_img * (1 - heatmap_alpha)
+    cv_out = np.clip(cv_out, a_min=0, a_max=255).astype(np.uint8)
+
+    return cv_out
+
+
+def get_heatmap_batch(img_batch, pred_batch):
+    imgrid = torchvision.utils.make_grid(img_batch).cpu()
+    magn_batch = torch.norm(pred_batch, p=2, dim=1, keepdim=True)
+    magngrid = torchvision.utils.make_grid(magn_batch)
+    magngrid = magngrid[0, :, :]
+    imgrid = unnormalize(imgrid).squeeze_()
+
+    cv_magn = magngrid.detach().cpu().numpy()
+    cv_img = imgrid.permute(1, 2, 0).detach().cpu().numpy()
+    cv_out = get_heatmap_cv(cv_img, cv_magn, max_flow_mag=9)
+    out = np.asarray(cv_out).astype(np.double) / 255.0
+
+    out = torch.from_numpy(out).permute(2, 0, 1)
+    return out
+
+
+def save_heatmap_cv(img, magn, path, max_flow_mag=7):
+    cv_out = get_heatmap_cv(img, magn, max_flow_mag)
+    out = Image.fromarray(cv_out)
+    out.save(path, quality=95)

weights/download_weights.sh

@@ -0,0 +1,2 @@
+wget https://www.dropbox.com/s/rb8zpvrbxbbutxc/global.pth?dl=0 -O ./weights/global.pth
+wget https://www.dropbox.com/s/pby9dhpr6cqziyl/local.pth?dl=0 -O ./weights/local.pth

weights/global.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aeaf8e135c018fb5298c0ed3d2fc47b39dd656708463f86713187ca62bb9586d
+size 82546417

weights/local.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4413d59e61f6ecb2960263385e3b17e248acd3d1db4575666c0a02d121c7aaa
+size 247495840