Added image processing

app.py

@@ -184,7 +184,7 @@ def extract_emb(image, modelName=FRMODELS[0], mode=None, imgPath=None):
         embs = F.normalize(embs, dim=1)
         embs = embs.detach().numpy()
         embs = embs.squeeze(0)
-        if mode != None:            
+        if mode != None: 
             subject = imgPath.split('/')[-2]
             os.makedirs(f'./embeddings/{subject}', exist_ok=True)
             emb_path = f'./embeddings/{subject}/{mode}-emb.txt'
@@ -227,14 +227,13 @@ with gr.Blocks() as demo:
         <p>This demo shows <strong>Suraksh.AI's</strong> biometric search solution under <strong>FHE</strong>.</p>
         <ul>
         <li><strong>Scenario 1</strong>: Searching an enrolled subject. For this scenario, the reference and probe should be from the same subject. Expected outcome: <span style='color: green; font-weight: bold;'>✔️ Found</span></li>
-        <li><strong>Scenario 2</strong>: Searching an enrolled subject with high recognition threshold. For this scenario, the reference and probe should be from the same subject and increase the recognition threshold. Expected outcome: <span style='color: red; font-weight: bold;'>❌ Not Found</span></li>
+        <li><strong>Scenario 2</strong>: Searching an enrolled subject with high recognition threshold. For this scenario, the reference and probe should be from the same subject and the recognition threshold set to a high value. Expected outcome: <span style='color: red; font-weight: bold;'>❌ Not Found</span></li>
         <li><strong>Scenario 3</strong>: Searching a non-enrolled subject. For this scenario, choose a probe not enrolled. Expected outcome: <span style='color: red; font-weight: bold;'>❌ Not Found</span></li>
-        <li><strong>Scenario 4</strong>: Searching a non-enrolled subject with low recognition threshold. For this scenario, choose a probe not enrolled and lower the recognition threshold. Expected outcome: <span style='color: green; font-weight: bold;'>✔️ Found</span></li>
+        <li><strong>Scenario 4</strong>: Searching a non-enrolled subject with low recognition threshold. For this scenario, choose a probe not enrolled and lower the high recognition threshold. Expected outcome: <span style='color: green; font-weight: bold;'>✔️ Found</span></li>
         </ul>
         """
     )
-
-
+    
 
     with gr.Row():
         gr.Markdown("### Setup Phase: 🔐 Generate the FHE public and secret keys.")
@@ -254,6 +253,7 @@ with gr.Blocks() as demo:
         gr.Markdown("## Phase 1: Enrollment")    
     with gr.Row():
         gr.Markdown("### Step 1: Upload or select a reference facial image for enrollment.")  
+
     with gr.Row():
         selectedImagePath = gr.State()
         count = gr.State(0)
@@ -277,8 +277,6 @@ with gr.Blocks() as demo:
                 label="Choose a face recognition model"
             )
         with gr.Column():
-            selectedImagePath = gr.State()
-            refDB_gallery.select(fn=get_selected_image, inputs=None, outputs=[image_input_enroll, selectedImagePath])
             key_button = gr.Button("Generate embedding")
             enroll_emb_text = gr.JSON(label="Reference embedding")
             mode = gr.State("enroll")
@@ -287,7 +285,7 @@ with gr.Blocks() as demo:
 
 
     with gr.Row():
-        gr.Markdown("""Facial embeddings are **INVERTIBLE** and lead to the **RECONSTRUCTION** of their raw facial images.""")   
+        gr.Markdown("""Facial embeddings are <span style='color:red; font-weight:bold'>INVERTIBLE</span> and lead to the <span style='color:red; font-weight:bold'>RECONSTRUCTION</span> of their raw facial images.""")   
     with gr.Row():
         gr.Markdown("### Example:")   
     with gr.Row():
@@ -299,7 +297,7 @@ with gr.Blocks() as demo:
         Reconstructed_image = gr.Image(label="Reconstructed")        
         btn.click(fn=load_rec_image, outputs=Reconstructed_image)
     with gr.Row():
-        gr.Markdown("""Facial embeddings protection is a must! At **Suraksh.AI**, we protect facial embeddings using FHE.""")
+        gr.Markdown("""Facial embeddings protection is a <span style='color:red; font-weight:bold'>MUST!</span> At **Suraksh.AI**, we protect facial embeddings using FHE.""")
 
     with gr.Row():
         gr.Markdown("### Step 3: 🔒 Encrypt reference embedding using FHE.")
@@ -354,18 +352,19 @@ with gr.Blocks() as demo:
             image_upload_auth = gr.Image(label="Upload a facial image.", type="pil", sources="upload")
             image_upload_auth.change(fn=crop_face_to_112x112, inputs=image_upload_auth, outputs=[image_input_auth,selectedImagePath_auth])
         with gr.Column():
+            # selectedImagePath_auth = gr.State()
             prob_gallery = gr.Gallery(value=example_images_auth, columns=3)
             prob_gallery.select(fn=get_selected_image_auth, inputs=None, outputs=[image_input_auth,selectedImagePath_auth])
         with gr.Column():
             selectedImage = gr.Image(type="pil", label="Probe facial image", interactive=False)
             image_input_auth.change(fn=lambda img: img, inputs=image_input_auth, outputs=selectedImage)
+
+
     
     with gr.Row():
         gr.Markdown("### Step 2: Generate probe facial embedding.") 
     with gr.Row():
-        with gr.Column():
-            selectedImagePath_auth = gr.State()
-            prob_gallery.select(fn=get_selected_image_auth, inputs=None, outputs=[image_input_auth,selectedImagePath_auth])
+        with gr.Column():            
             key_button = gr.Button("Generate embedding")
             prob_emb_text = gr.JSON(label="Probe embedding")
             mode = gr.State("auth")