File size: 2,608 Bytes
5333123
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
import gradio as gr

from timeit import default_timer as timer
from typing import Tuple , Dict
import tensorflow as tf 
import numpy as np

from PIL import Image
import os

# 1.Import and class names setup
class_names = ['CNV','DME','DRUSEN','NORMAL']


# 2. Model annd transforms prepration
# model = tf.keras.models.load_model(
#     'oct_classification_final_model_lg.keras', custom_objects=None, compile=True, safe_mode=True
# )
model = tf.keras.models.load_model(
    'oct_classification_final_model_lg.keras', custom_objects=None, compile=True, safe_mode=False
)


# Load save weights

# 3.prediction function (predict())

def load_and_prep_imgg(img : Image.Image, img_shape=224, scale=True):
    # if not isinstance(filename, str):
    #     raise ValueError("The filename must be a string representing the file path.")
    # img = tf.io.read_file(filename)
    # img = tf.io.decode_image(img, channels=3)
    # img = tf.image.resize(img, size=[img_shape, img_shape])
    # if scale:
    #     return img / 255
    # else:
    #     return img
    img = img.resize((img_shape, img_shape))
    img = np.array(img)
    if img.shape[-1] == 1:  # If the image is grayscale
        img = np.stack([img] * 3, axis=-1)
    img = tf.convert_to_tensor(img, dtype=tf.float32)
    if scale:
        return img / 255.0
    else:
        return img
        
def predict(img) -> Tuple[Dict,float] :

  start_time = timer()

  image =  load_and_prep_imgg(img)
  #image = Image.open(image)  

  pred_img = model.predict(tf.expand_dims(image, axis=0))
  pred_class = class_names[pred_img.argmax()]
  print(f"Predicted macular diseases is: {pred_class} with probability: {pred_img.max():.2f}")

  

  
  end_time = timer()
  pred_time = round(end_time - start_time , 4)

  return pred_class , pred_time

### 4. Gradio app - our Gradio interface + launch command

title = 'Macular Disease Classification'
description = 'Feature Extraction VGG model to classify Macular Diseases by OCT'
article = 'Created with TensorFlow Model Deployment'
# Create example list

example_list = [['examples/'+ example] for example in os.listdir('examples')]
example_list

# create a gradio demo
demo = gr.Interface(fn=predict ,
                    inputs=gr.Image(type='pil'),
                    outputs=[gr.Label(num_top_classes = 3 , label= 'prediction'),
                            gr.Number(label= 'Prediction time (s)')],
                    examples = example_list,
                    title = title,
                    description = description,
                    article= article)

# Launch the demo
demo.launch(debug= False)