tumor seg and chest xray added

.gitattributes

@@ -34,3 +34,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
 model/variables/variables.data-00000-of-00001 filter=lfs diff=lfs merge=lfs -text
+segment_model/V2/variables/variables.data-00000-of-00001 filter=lfs diff=lfs merge=lfs -text
+xray_resnet_model/variables/variables.data-00000-of-00001 filter=lfs diff=lfs merge=lfs -text

app.py

@@ -1,20 +1,60 @@
 import gradio as gr
-from predict import makepredictions
+from predict import makepredictions, xray_predict
+from tumorseg import make_segmentation
 
 with gr.Blocks() as demo:
-    gr.Markdown("Brain Tumor Image classification")
-    gr.Markdown("We detect 4 Types of Tumor with 99.12% Accuracy:")
-    gr.Markdown("Glioma  |   Meningioma  |   No Tumor  |   Pituitary")
-    image_input = gr.Image(label='Tumor Image', type='filepath', height=400)
-    with gr.Row():
-        image_button = gr.Button("Predict")
-        y_pred = output = gr.Textbox(label="Tumor Type:")
-        image_button.click(makepredictions, inputs=image_input, outputs=y_pred)
+    with gr.Tab('Brain Tumor | Tumor Segmentation'):
+        gr.Markdown("""
+            <div align='center'>Tumor segmentations
+            </div>
+            <div align='center'>
+                   > You could segment you mri image. I hope it will find you well       
+            </div>
+             """)
+        with gr.Row():
+            image_input = gr.Image(label='Tumor Image', type='filepath', height=480, interactive=True)
+            output_image = gr.Image(label='label Segmented Image', height=480, interactive=False)
+        image_button = gr.Button("Segment image")
+        image_button.click(make_segmentation, inputs=image_input, outputs=output_image)
+
+    with gr.Tab('Brain Tumor | Tumor Classification'):
+        gr.Markdown("""
+        <div align='center'>
+        Brain Tumor Image classification
+        </div>
+        <div align='center'>
+            >We detect 4 Types of Tumor with 99.12% Accuracy:
+            >Glioma  |   Meningioma  |   No Tumor  |   Pituitary
+        """)
+        image_input = gr.Image(label='Tumor Image', type='filepath', height=480)
+        with gr.Row():
+            image_button = gr.Button("Predict")
+            y_pred = gr.Textbox(label="Tumor Type:")
+            image_button.click(makepredictions, inputs=image_input, outputs=y_pred)
+
+    with gr.Tab('Chest X-ray | COVID19, NORMAL, PNEUMONIA,TURBERCULOSIS'):
+        gr.Markdown("""
+                    <div align='center'>Chest X-ray classifications
+                    </div>
+                    <div align='center'>
+                       > We COVID19, NORMAL, PNEUMONIA,TURBERCULOSIS from your chest x-ray image
+                    </div>
+                          """)
+
+        image_input = gr.Image(label='Chest X-ray', type='filepath', height=480)
+        with gr.Row():
+            image_button = gr.Button("Predict")
+            y_pred = gr.Textbox(label="Disease :")
+            image_button.click(xray_predict, inputs=image_input, outputs=[y_pred])
+
     gr.Markdown(
         """
-           This model is trained on the Kaggle dataset:
-                  1. Only for the educational purpose
-                  2. Play with that and Have fun.
+        <div align='center'>
+           These model is trained on the Kaggle dataset:
+           </div>
+           <div align='center'>
+                  > Only for the educational purpose
+                  > Play with that and Have fun.
+            </div>      
         """)
-
 demo.launch()

predict.py

@@ -1,6 +1,7 @@
 import tensorflow as tf
 import tf_keras
 model_mri = tf_keras.models.load_model('model')
+model_xray = tf_keras.models.load_model('xray_resnet_model')
 
 def load_image_with_path(path):
     img = tf.io.read_file(path)
@@ -24,3 +25,13 @@ def makepredictions(path):
         a = "Result : Pituitary Tumor"
     return a
     # {'glioma': 0, 'meningioma': 1, 'notumor': 2, 'pituitary': 3}
+
+def xray_predict(path):
+    img = load_image_with_path(path)
+    predications = model_xray.predict(tf.expand_dims(img, axis=0))
+    a = int(tf.argmax(tf.squeeze(predications)))
+    xray_classes = ['COVID19','NORMAL', 'PNEUMONIA',' TURBERCULOSIS']
+    a = xray_classes[a]
+
+    return a
+    # {'COVID19': 0, 'NORMAL': 1, 'PNEUMONIA': 2, 'TURBERCULOSIS': 3}

segment_model/V2/fingerprint.pb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e6526424daeeaa19ccabb7aceafada7ff465e4b877fe0babed162fd3dda1693
+size 56

segment_model/V2/info.txt

@@ -0,0 +1,2 @@
+binary accuracy:99%
+best model till yet : 20-03-2024

segment_model/V2/keras_metadata.pb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:65a8d82887d611745ad6fbeecae6944a05f6996faeb3cc6997e527ace88e84e7
+size 125191

segment_model/V2/saved_model.pb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e3b802b0e98405daeb5e53dc69cd72746ebbf8aa05d663fbcf6335916a818cf1
+size 1158923

segment_model/V2/variables/variables.data-00000-of-00001

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbb9c2edd0e2cef3a589fb44c45dc9bc923acbbb7b5239c2832b1b79a4bcdb86
+size 372540402

tumorseg.py

@@ -0,0 +1,73 @@
+import os
+import tensorflow as tf
+from tensorflow.keras import backend as K
+import tf_keras
+import numpy as np
+import matplotlib.pyplot as plt
+import matplotlib
+matplotlib.use('agg')
+def dice_coefficients(y_true, y_pred, smooth=100):
+    y_true_flatten = K.flatten(y_true)
+    y_pred_flatten = K.flatten(y_pred)
+
+    intersection = K.sum(y_true_flatten * y_pred_flatten)
+    union = K.sum(y_true_flatten) + K.sum(y_pred_flatten)
+    return (2 * intersection + smooth) / (union + smooth)
+
+
+def dice_coefficients_loss(y_true, y_pred, smooth=100):
+    return 1.0 - dice_coefficients(y_true, y_pred, smooth)
+
+
+def iou(y_true, y_pred, smooth=100):
+    intersection = K.sum(y_true * y_pred)
+    sum = K.sum(y_true + y_pred)
+    iou = (intersection + smooth) / (sum - intersection + smooth)
+    return iou
+
+
+def jaccard_distance(y_true, y_pred):
+    y_true_flatten = K.flatten(y_true)
+    y_pred_flatten = K.flatten(y_pred)
+    return -iou(y_true_flatten, y_pred_flatten)
+
+segmodel = tf_keras.models.load_model("segment_model/V2", custom_objects={'dice_coefficients_loss': dice_coefficients_loss, 'iou': iou, 'dice_coefficients': dice_coefficients  } )
+
+def load_image_for_pred(image_path):
+    img = tf.keras.utils.load_img(
+        image_path,
+        color_mode='rgb',
+        target_size=(256, 256),
+        interpolation='nearest',
+        keep_aspect_ratio=False
+    )
+    img = tf.keras.utils.img_to_array(img) / 255
+    return np.array([img])
+
+def make_segmentation(image_path):
+    img = load_image_for_pred(image_path)
+    predicted_img = segmodel.predict(img)
+
+    plt.figure(figsize=(5, 3))
+    plt.subplot(1, 3, 1)
+    plt.imshow(np.squeeze(img))
+    plt.title('Original Image')
+    plt.axis(False)
+
+
+    plt.subplot(1, 3, 2)
+    plt.imshow(np.squeeze(predicted_img) > 0.5)
+    plt.title('Prediction')
+    plt.axis(False)
+
+    plt.subplot(1, 4, 4)
+    plt.imshow(np.squeeze(img))
+    plt.imshow(np.squeeze(predicted_img) > 0.5, cmap='gray', alpha=0.5)
+    plt.title('Image w/h Mask')
+    plt.axis(False)
+    save_file_name = os.path.splitext(image_path)[0] + '_segmented.png'
+    plt.savefig(save_file_name)
+
+    return save_file_name
+
+

xray_resnet_model/fingerprint.pb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ae9236c9c9287ca25c24110c97d624e0b57bb8961c8ef20f435c0c00cd6124f
+size 57

xray_resnet_model/info.txt

@@ -0,0 +1,2 @@
+accuracy:93%
+best model : 20-03-2024

xray_resnet_model/keras_metadata.pb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3b251b17e58a01b8d485937055ac98409b680bcc5c126f977da35807827838ba
+size 605636

xray_resnet_model/saved_model.pb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:51fec9fb52f34de9872bac4e8c20db741ed1b286b8cc4329b483690a1f82243d
+size 3067953

xray_resnet_model/variables/variables.data-00000-of-00001

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a1fc2f8c15c651684e0d95d2e94d1367500afb712778a2dc9d6b6bea4d8eb289
+size 216397949