Added predictions mapper imfing#2

app.py

@@ -1,8 +1,11 @@
+import json
 import os
 import sys
 
 # Flask
+from PIL import Image
 from flask import Flask, redirect, url_for, request, render_template, Response, jsonify, redirect
+from keras_applications import imagenet_utils
 from werkzeug.utils import secure_filename
 from gevent.pywsgi import WSGIServer
 
@@ -12,51 +15,57 @@
 
 from keras.applications.vgg16 import preprocess_input
 from tensorflow.keras.models import load_model
-from tensorflow.keras.preprocessing import image
-from keras.backend import backend as K
+from keras import backend as K
 
 K.clear_session()
 # Some utilites
 import numpy as np
 from util import base64_to_pil
 
-
 # Declare a flask app
 app = Flask(__name__)
 
+TARGET_CLASSES = {
+    0: "Normal",
+    1: "Tuberculosis"
 
-
+}
 
 print('Model loaded. Check http://127.0.0.1:5000/')
 
-
+graph = tf.get_default_graph()
 # Model saved with Keras model.save()
 MODEL_PATH = 'models/model.h5'
 
 # Load your own trained model
 model = load_model(MODEL_PATH)
-model._make_predict_function()          # Necessary
+model._make_predict_function()  # Necessary
 print('Model loaded. Start serving...')
 
 
-def model_predict(img, model):
-    img = img.resize((96, 96))
-
+def model_predict(img):
     # Preprocessing the image
-    x = image.img_to_array(img)
-    # x = np.true_divide(x, 255)
-    img = np.reshape(img, [1,96, 96, 3])
-    # x = np.expand_dims(x, axis=0)
-    #
-    # # Be careful how your trained model deals with the input
-    # # otherwise, it won't make correct prediction!
-    # x = preprocess_input(x, mode='tf')
+    img = prepare_image(img)
+    preds = []
+    with tf.Session(graph=graph) as sess:
+        sess.run(tf.global_variables_initializer())
+
+        preds = model.predict(img)
 
+    return preds
 
 
+def prepare_image(image):
+    # if the image mode is not RGB, convert it
+    if image.mode != "RGB":
+        image = image.convert("RGB")
 
-    preds = model.predict(img)
-    return preds
+    im = image.resize((96, 96), Image.ANTIALIAS)
+    doc = keras.preprocessing.image.img_to_array(im)  # -> numpy array
+
+    doc = np.expand_dims(doc, axis=0)
+
+    return doc
 
 
 @app.route('/', methods=['GET'])
@@ -67,32 +76,28 @@ def index():
 
 @app.route('/predict', methods=['GET', 'POST'])
 def predict():
+    prediction = {"Message": "Cannot process"}
     if request.method == 'POST':
-        # Get the image from post request
-        img = base64_to_pil(request.json)
 
-        # Save the image to ./uploads
-        # img.save("./uploads/image.png")
+        try:
+
+            # Get the image from post request
+            img = base64_to_pil(request.json)
 
-        # Make prediction
-        preds = model_predict(img, model)
-        print(preds)
+            # Make prediction
+            preds = model_predict(img)
+            index_min = np.argmin(preds[0])
 
-        # Process your result for human
-        # pred_proba = "{:.3f}".format(np.amax(preds))    # Max probability
-        # pred_class = decode_predictions(preds, top=1)   # ImageNet Decode
+            return jsonify(result=TARGET_CLASSES[index_min], probability=json.dumps(str(preds[0][index_min])))
 
-        # result = str(pred_class[0][0][1])               # Convert to string
-        # result = result.replace('_', ' ').capitalize()
-        #
-        # # Serialize the result, you can add additional fields
-        # return jsonify(result=result, probability=pred_proba)
+        except Exception as ex:
+            return None
 
     return None
 
 
 if __name__ == '__main__':
-    # app.run(port=5002, threaded=False)
+    # app.run(port=5000, threaded=False, debug=False)
 
     # Serve the app with gevent
     http_server = WSGIServer(('0.0.0.0', 5000), app)