Update app.py

app.py

@@ -1,100 +1,30 @@
 import torch
+import cv2
 import gradio as gr
 from ultralytics import YOLO
+import numpy as np
 from huggingface_hub import hf_hub_download
 
-# 🌊 Model Setup
+# Download and load model
 model_path = hf_hub_download(repo_id="mohamedrayyan/murakamodel", filename="best.pt")
 model = YOLO(model_path)
 
-# 🐚 Detection Function
+# Detection function with confidence threshold
 def detect_objects(image, confidence_threshold):
-    results = model(image, conf=confidence_threshold)
+    results = model(image, conf=confidence_threshold)  # Use confidence threshold
     annotated_frame = results[0].plot()
-    
-    # Add custom overlay
-    annotated_frame = cv2.putText(
-        annotated_frame, 
-        "🪸 Coral Health Assessment", 
-        (10, 30), 
-        cv2.FONT_HERSHEY_SIMPLEX, 
-        0.7, 
-        (0, 150, 255), 
-        2
-    )
     return annotated_frame
 
-# 🌴 UI Design
-with gr.Blocks(theme=gr.themes.Soft(primary_hue="teal")) as demo:
-    # Header Section
-    with gr.Row(variant="compact"):
-        gr.HTML("""
-        <div style="text-align: center">
-            <h1 style="color: #2E86AB">🌊 Muraka: Coral AI Doctor 🐚</h1>
-            <p style="color: #4A7C59">Advanced Coral Health Detection System</p>
-        </div>
-        """)
-
-    # Main Interface
-    with gr.Row():
-        with gr.Column(scale=1):
-            gr.Markdown("### 🐡 Upload Coral Image")
-            input_image = gr.Image(type="numpy", label="Drag & Drop Reef Image")
-            
-            with gr.Accordion("⚙️ Analysis Settings", open=False):
-                confidence_slider = gr.Slider(
-                    minimum=0.1, 
-                    maximum=1.0, 
-                    value=0.5, 
-                    label="Detection Sensitivity",
-                    info="Lower = More Sensitive | Higher = More Conservative"
-                )
-                gr.Markdown("💡 **Pro Tip:** Start with 0.5 sensitivity for balanced results!")
-            
-            with gr.Row():
-                submit_button = gr.Button("🔍 Analyze", variant="primary")
-                clear_button = gr.Button("🧹 Clear", variant="secondary")
-
-        with gr.Column(scale=2):
-            gr.Markdown("### 📋 Detection Results")
-            output_image = gr.Image(type="numpy", label="Health Assessment")
-            
-            with gr.Group(visible=False) as stats_box:
-                gr.Markdown("**📊 Quick Statistics**")
-                healthy_count = gr.Number(label="Healthy Colonies")
-                bleached_count = gr.Number(label="Bleached Areas")
-            
-            download_button = gr.Button("💾 Download Report", variant="primary")
-
-    # Interactive Functions
-    def update_stats(results):
-        counts = {"healthy": 0, "bleached": 0}
-        for box in results[0].boxes:
-            if model.names[int(box.cls)] == "healthy":
-                counts["healthy"] += 1
-            else:
-                counts["bleached"] += 1
-        return counts["healthy"], counts["bleached"]
-
-    submit_button.click(
-        fn=detect_objects,
-        inputs=[input_image, confidence_slider],
-        outputs=output_image
-    ).then(
-        fn=lambda: gr.Group(visible=True),
-        outputs=stats_box
-    )
-
-    clear_button.click(
-        lambda: [None, None, gr.Group(visible=False)],
-        outputs=[input_image, output_image, stats_box]
-    )
-
-    download_button.click(
-        fn=lambda x: (x, "coral_report.png"),
-        inputs=output_image,
-        outputs=gr.File(label="Download Report")
-    )
-
-# 🚀 Launch
-demo.launch()
+# Create Gradio interface with slider
+gr.Interface(
+    fn=detect_objects,
+    inputs=[
+        gr.Image(type="numpy", label="Upload Coral Image"),
+        gr.Slider(0.0, 1.0, value=0.25, label="Confidence Threshold", 
+                 info="Lower = more sensitive, Higher = fewer false positives")
+    ],
+    outputs=gr.Image(type="numpy", label="Analysis Results"),
+    title="🪸 Muraka - AI Doctor for Corals",
+    description="Upload coral images to assess health status. Adjust confidence threshold to control detection sensitivity.",
+    allow_flagging="never"
+).launch()