slicing input

app.py

@@ -79,15 +79,15 @@ class Detection:
             device='cuda:0'
         )
 
-    def detect_from_image(self, image):
+    def detect_from_image(self, image,slice_width_input,slice_height_input,overlap_width_input,overlap_height_input):
         # Perform sliced prediction with both models
         results1 = get_sliced_prediction(
             image=image,
             detection_model=self.model1,
-            slice_height=256,
-            slice_width=256,
-            overlap_height_ratio=0.5,
-            overlap_width_ratio=0.5,
+            slice_height=slice_height_input,
+            slice_width=slice_width_input,
+            overlap_height_ratio=overlap_height_input,
+            overlap_width_ratio=overlap_width_input,
             postprocess_type='NMS',
             postprocess_match_metric='IOU',
             postprocess_match_threshold=0.1,
@@ -97,10 +97,10 @@ class Detection:
         results2 = get_sliced_prediction(
             image=image,
             detection_model=self.model2,
-            slice_height=256,
-            slice_width=256,
-            overlap_height_ratio=0.5,
-            overlap_width_ratio=0.5,
+            slice_height=slice_height_input,
+            slice_width=slice_width_input,
+            overlap_height_ratio=overlap_height_input,
+            overlap_width_ratio=overlap_width_input,
             postprocess_type='NMS',
             postprocess_match_metric='IOU',
             postprocess_match_threshold=0.1,
@@ -321,13 +321,13 @@ def upload_image(image):
     return image
 
 @spaces.GPU
-def apply_detection(image):
+def apply_detection(image,slice_width_input,slice_height_input,overlap_width_input,overlap_height_input):
     """Run object detection on the uploaded image and return the annotated image.""" 
     # Convert image from PIL to NumPy array
     img = np.array(image)
 
     # Perform detection and get COCO annotations
-    annotations = detection.detect_from_image(img)
+    annotations = detection.detect_from_image(img,slice_width_input,slice_height_input,overlap_width_input,overlap_height_input)
 
     # Draw the annotations on the image using OpenCV
     annotated_image = detection.draw_annotations(img, annotations)
@@ -614,6 +614,19 @@ with gr.Blocks() as demo:
                     
         """)
 
+    with gr.Row(visible=False) as slicing_text:
+        gr.Markdown("### Choose the width and height dimension and the overlapping ratio of the slice to determine how small the model can detect")
+
+
+    with gr.Row(visible=False) as slicing_dim_input:
+        # Add inputs for width and height
+        slice_width_input = gr.Number(label="Slice Width (pixels)", value=256)
+        slice_height_input = gr.Number(label="Slice Height (pixels)", value=256)
+
+    with gr.Row(visible=False) as slicing_overlap_input:
+        overlap_width_input = gr.Slider(0, 1, step=0.01, label="Overlap Width Ratio", value=0.5)
+        overlap_height_input = gr.Slider(0, 1, step=0.01, label="Overlap Height Ratio", value=0.5)
+    
 
     with gr.Row(visible=False) as input_row:
         # Image Upload and Display in two columns
@@ -626,7 +639,7 @@ with gr.Blocks() as demo:
             output_image_component = gr.Image(type="pil", label="Annotated Image")
             apply_detection_btn = gr.Button("Apply Detection", variant='primary')
             output_annotations = gr.State()  # Store annotations
-            apply_detection_btn.click(apply_detection, inputs=upload_image_component, outputs=[output_image_component, output_annotations])
+            apply_detection_btn.click(apply_detection, inputs=[upload_image_component,slice_width_input,slice_height_input,overlap_width_input,overlap_height_input], outputs=[output_image_component, output_annotations])
 
 
 
@@ -753,6 +766,9 @@ with gr.Blocks() as demo:
                     ).then(
                         lambda login_state: (
                             gr.update(visible=login_state),  # Show header_row
+                            gr.update(visible=login_state),  # Show slicing text
+                            gr.update(visible=login_state),  # Show slicing_dim_input
+                            gr.update(visible=login_state),  # Show slicing_overlap_input
                             gr.update(visible=login_state),  # Show input_row
                             gr.update(visible=login_state),  # Show area_graph_row
                             gr.update(visible=login_state),  # Show area_btn_row
@@ -768,7 +784,10 @@ with gr.Blocks() as demo:
                             gr.update(visible=not login_state) # for login
                         ),
                         inputs=login_successful,
-                        outputs=[header_row, 
+                        outputs=[header_row,
+                                 slicing_text,
+                                 slicing_dim_input,
+                                 slicing_overlap_input,
                                  input_row, 
                                  area_graph_row, 
                                  area_btn_row,