fixed color conversion

app.py

@@ -29,8 +29,6 @@ import boto3
 import os
 os.environ['KMP_DUPLICATE_LIB_OK'] = 'True'
 os.environ['OMP_NUM_THREADS'] = '4'
-os.environ['AWS_ACCESS_KEY_ID'] = 'AKIA3JAMX4K53MFDKMGJ'
-os.environ['AWS_SECRET_ACCESS_KEY'] = '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'
 
 app_version = 'dsdg_vid_2'
 
@@ -97,7 +95,7 @@ def extract_face(img):
     face = None
     if img is None:
         return face
-    grey = cv.cvtColor(img, cv.COLOR_RGB2GRAY)
+    grey = cv.cvtColor(img, cv.COLOR_BGR2GRAY)
     faces = faceClassifier.detectMultiScale(
         grey, scaleFactor=1.1, minNeighbors=4)
     if len(faces):
@@ -161,14 +159,14 @@ def prepare_data_dsdg(images, boxes, depths):
     for i, (image, bbox, depth_img) in enumerate(
             zip(images, boxes, depths)):
         x, y, x2, y2 = bbox
-        depth_img = cv.cvtColor(depth_img, cv.COLOR_RGB2GRAY)
+        depth_img = cv.cvtColor(depth_img, cv.COLOR_BGR2GRAY)
         image = image[y:y2, x:x2]
         depth_img = depth_img[y:y2, x:x2]
 
         image_x[i, :, :, :] = cv.resize(image, (256, 256))
         # transform to binary mask --> threshold = 0 
         depth_x[i, :, :] = cv.resize(depth_img, (32, 32))
-    image_x = image_x.transpose((0, 3, 1, 2))
+    image_x = image_x[:, :, :, ::-1].transpose((0, 3, 1, 2))
     image_x = transform(image_x)
     image_x = torch.from_numpy(image_x.astype(float)).float()
     depth_x = torch.from_numpy(depth_x.astype(float)).float()
@@ -235,7 +233,6 @@ def process_video(vid_path, dsdg_thresh):
         # Process only every 5th frame
         if frame_counter % 5 == 0:
             # Run inference on the current frame
-            frame = cv.cvtColor(frame, cv.COLOR_BGR2RGB)
             img, bbox, depth_img = analyze_face(frame)
             if bbox and (depth_img is not None):
                 inference_images.append(img)
@@ -254,7 +251,7 @@ def process_video(vid_path, dsdg_thresh):
         w = x2 - x
         h = y2 - y
         frame_cls = 'Real' if score >= dsdg_thresh else 'Spoof'
-        color_dsdg = (0, 255, 0) if frame_cls == 'Real' else (255, 0, 0)
+        color_dsdg = (0, 255, 0) if frame_cls == 'Real' else (0, 0, 255)
         text = f'{cls_dsdg} {w}*{h}'
         cv.rectangle(img, (x, y), (x2, y2), color_dsdg, 2)
         cv.putText(img, text, (x, y2 + 30), cv.FONT_HERSHEY_COMPLEX, 1, color_dsdg)
@@ -262,8 +259,7 @@ def process_video(vid_path, dsdg_thresh):
     out_dsdg = cv.VideoWriter(output_vid_path, fourcc, 6.0, (input_width, input_height))
     for img in all_frames:
         # Write the DSDG frame to the output video
-        img_dsdg = cv.cvtColor(img, cv.COLOR_RGB2BGR)
-        out_dsdg.write(img_dsdg)
+        out_dsdg.write(img)
     out_dsdg.release()
     text_dsdg = f'Label: {cls_dsdg}, average real confidence: {res_dsdg}\nFrames used: {len(scores)}\nConfidences: {scores}'
     return vid_path, {'Not supported right now': 0}, -1, output_vid_path, text_dsdg, res_dsdg