Update app.py

app.py

@@ -48,15 +48,15 @@ def dice_coef(y_true, y_pred, smooth=1):
 # Load the model
 seg_model = keras.models.load_model('seg_unet_model.h5', custom_objects={'Combo_loss': Combo_loss, 'dice_coef': dice_coef})
 
-inputs = gr.inputs.Image(type="pil", label="Upload an image")
+# inputs = gr.inputs.Image(type="pil", label="Upload an image")
 image_output = gr.outputs.Image(type="pil", label="Output Image")
 # outputs = gr.outputs.HTML() #uncomment for single class output 
 
 
 def gen_pred(img, model=seg_model):
-    pil_image = img.convert('RGB')
-    open_cv_image = np.array(pil_image)
-    img = open_cv_image[:, :, ::-1].copy() 
+    # pil_image = img.convert('RGB')
+    # open_cv_image = np.array(pil_image)
+    # img = open_cv_image[:, :, ::-1].copy() 
     # # img = cv2.imread("./003e2c95d.jpg")
     img = img[::IMG_SCALING[0], ::IMG_SCALING[1]]
     img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
@@ -64,7 +64,8 @@ def gen_pred(img, model=seg_model):
     img = tf.expand_dims(img, axis=0)
     pred = model.predict(img)
     pred = np.squeeze(pred, axis=0)
-    pil_img = Image.fromarray(pred, 'RGB')
+    pil_img = Image.fromarray((x * 255).astype(np.uint8))
+    # pil_img = Image.fromarray(pred, 'RGB')
     # im = Image.fromarray((pred * 255).astype(np.uint8))
     # img_bytes = pred.tobytes()
     # nparr = np.frombuffer(img_bytes, np.byte)
@@ -77,7 +78,7 @@ description = "Upload an image and get prediction mask"
 # css_code='body{background-image:url("file=wave.mp4");}'
 
 gr.Interface(fn=gen_pred, 
-             inputs=[gr.Image(type='pil')], 
+             inputs=[gr.Image(type='numpy')], 
              outputs=gr.Image(type='numpy'), 
              title=title, 
              examples=[["003e2c95d.jpg"], ["003b50a15.jpg"], ["003b48a9e.jpg"], ["0038cbe45.jpg"], ["00371aa92.jpg"]],