Update app.py

app.py

@@ -230,28 +230,54 @@ desc = (
     "If both classifiers fire, the stronger probability is chosen (fallback). Thresholds adjustable."
 )
 
+examples_df = pd.DataFrame({
+    "Image": [
+        "images/NORMAL.jpeg",
+        "images/VIRAL.jpeg",
+        "images/BACT.jpeg",
+    ],
+    "Label": ["NORMAL", "VIRAL", "BACTERIAL"]
+})
+
 with gr.Blocks(title=title) as demo:
+
     gr.Markdown(f"### {title}")
     gr.Markdown(desc)
 
     with gr.Row():
-        with gr.Column():
+        with gr.Column(scale=1):
+
             image_input = gr.Image(
                 type="numpy",
-                label="Upload chest X-ray (RGB or grayscale)"
+                label="Upload Chest X-ray"
+            )
+
+            thresh_b = gr.Slider(
+                minimum=0.1, maximum=0.9, step=0.01, value=0.5,
+                label="Bacterial Threshold"
+            )
+
+            thresh_v = gr.Slider(
+                minimum=0.1, maximum=0.9, step=0.01, value=0.5,
+                label="Viral Threshold"
+            )
+
+            seg_thresh = gr.Slider(
+                minimum=0.1, maximum=0.9, step=0.01, value=0.5,
+                label="Segmentation Mask Threshold"
             )
-            thresh_b = gr.Slider(0.1, 0.9, value=0.5, step=0.01, label="Bacterial threshold (thresh_b)")
-            thresh_v = gr.Slider(0.1, 0.9, value=0.5, step=0.01, label="Viral threshold (thresh_v)")
-            seg_thresh = gr.Slider(0.1, 0.9, value=0.5, step=0.01, label="Segmentation mask threshold (seg_thresh)")
 
-            submit_btn = gr.Button("Run Inference")
+            with gr.Row():
+                clear_btn = gr.Button("Clear", variant="secondary")
+                submit_btn = gr.Button("Submit", variant="primary")
 
-        with gr.Column():
+
+        with gr.Column(scale=1):
             pred_label = gr.Label(num_top_classes=1, label="Prediction")
             prob_b = gr.Number(label="Bacterial Probability")
             prob_v = gr.Number(label="Viral Probability")
-            masked_img = gr.Image(type="pil", label="Masked Image (input × mask)")
-            seg_overlay = gr.Image(type="pil", label="Segmentation Overlay (red mask)")
+            masked_img = gr.Image(type="pil", label="Masked Image")
+            seg_overlay = gr.Image(type="pil", label="Segmentation Overlay")
 
     submit_btn.click(
         fn=inference_pipeline,
@@ -259,16 +285,35 @@ with gr.Blocks(title=title) as demo:
         outputs=[pred_label, prob_b, prob_v, masked_img, seg_overlay]
     )
 
-    gr.Examples(
-        examples=[
-            ["images/NORMAL.jpeg"],
-            ["images/VIRAL.jpeg"],
-            ["images/BACT.jpeg"],
-        ],
-        inputs=image_input,
-        label="Try Examples"
+    clear_btn.click(
+        fn=lambda: (None, None, None, None, None),
+        inputs=None,
+        outputs=[pred_label, prob_b, prob_v, masked_img, seg_overlay]
     )
 
+    with gr.Accordion("Try Examples", open=False):
+
+        examples_table = gr.Dataframe(
+            value=examples_df,
+            headers=["Image", "Label"],
+            datatype=["str", "str"],
+            interactive=False,
+            wrap=True,
+            height=300
+        )
+
+        load_btn = gr.Button("Load Selected Example")
+
+        def load_example(example_row):
+            if example_row is None or len(example_row) == 0:
+                return None
+            img_path = example_row[0]["Image"]
+            return img_path
+
+        load_btn.click(
+            fn=load_example,
+            inputs=examples_table,
+            outputs=image_input
+        )
 if __name__ == "__main__":
     demo.launch(share=False)
-