Update app.py

app.py

@@ -10,8 +10,8 @@ config = DetrConfig.from_pretrained("facebook/detr-resnet-50")
 model = DetrForObjectDetection.from_pretrained("facebook/detr-resnet-50", config=config)
 image_processor = DetrImageProcessor.from_pretrained("facebook/detr-resnet-50")
 
-# Initialize the pipeline
-od_pipe = pipeline(task='object-detection', model=model, image_processor=image_processor)
+# Initialize the pipeline, adjust the confidence_threshold if possible
+od_pipe = pipeline(task='object-detection', model=model, image_processor=image_processor, confidence_threshold=0.5)
 
 def draw_detections(image, detections):
     # Convert PIL image to a numpy array
@@ -32,7 +32,8 @@ def draw_detections(image, detections):
         # Draw rectangles and text with a larger font
         cv2.rectangle(np_image, (x_min, y_min), (x_max, y_max), (0, 255, 0), 2)
         label_text = f'{label} {score:.2f}'
-        cv2.putText(np_image, label_text, (x_min, y_min - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.8, (255, 255, 255), 2)
+        # Increase the font size and text thickness
+        cv2.putText(np_image, label_text, (x_min, y_min - 10), cv2.FONT_HERSHEY_SIMPLEX, 1.5, (255, 255, 255), 4)
     
     # Convert BGR to RGB for displaying
     final_image = cv2.cvtColor(np_image, cv2.COLOR_BGR2RGB)
@@ -40,10 +41,10 @@ def draw_detections(image, detections):
     return final_pil_image
 
 def get_pipeline_prediction(pil_image):
-    # Ensure the image is a PIL Image as expected by the model pipeline
-    if not isinstance(pil_image, Image.Image):
-        pil_image = Image.fromarray(pil_image.astype('uint8'), 'RGB')
     try:
+        # Ensure PIL image is passed correctly
+        if isinstance(pil_image, np.ndarray):
+            pil_image = Image.fromarray(pil_image.astype('uint8'), 'RGB')
         pipeline_output = od_pipe(pil_image)
         processed_image = draw_detections(pil_image, pipeline_output)
         return processed_image, pipeline_output
@@ -55,7 +56,7 @@ def get_pipeline_prediction(pil_image):
 with gr.Blocks() as demo:
     with gr.Row():
         with gr.Column():
-            inp_image = gr.Image(label="Input image", type="pil")  # Ensure the input is loaded as a PIL Image
+            inp_image = gr.Image(label="Input image")
             btn_run = gr.Button('Run Detection')
         with gr.Column():
             with gr.Tab("Annotated Image"):