File size: 3,486 Bytes
0d8c7b9
 
 
 
b2a9ce9
0d8c7b9
 
 
 
39b64b3
34f59b7
0d8c7b9
34f59b7
0d8c7b9
 
34f59b7
 
 
 
0d8c7b9
 
 
 
 
34f59b7
 
0d8c7b9
 
 
 
 
 
34f59b7
0d8c7b9
b2a9ce9
0d8c7b9
 
 
 
 
 
 
 
fc278cd
0d8c7b9
 
 
fc278cd
0d8c7b9
 
 
 
 
34f59b7
0d8c7b9
 
 
 
 
 
 
 
 
 
 
 
 
 
34f59b7
0d8c7b9
 
 
 
59ae286
34f59b7
 
 
 
 
 
 
 
0d8c7b9
 
 
 
 
c63e7d7
6794089
0d8c7b9
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
import gradio as gr

def create_demo():
    title = "πŸ”¬ EndoSight AI - Development Preview"
    
    description = """
## Advanced Gastrointestinal Polyp Detection & Analysis System

**Performance Metrics:**
- **88% mAP\\@0.5** detection accuracy
- **75.6% Dice Score** segmentation accuracy  
- **35+ FPS** real-time processing
- **YOLOv8 + U-Net** dual architecture with intelligent stabilization

### πŸ₯ Clinical Capabilities
- Real-time polyp detection and localization with smooth tracking
- Precise boundary segmentation with fluid heatmap visualization
- Intelligent measurement system with anti-fluctuation technology
- Automated size classification and risk assessment
- Optimized for endoscopy workflow integration

### 🎯 Technical Highlights
- **Multi-modal AI**: Combined object detection and segmentation
- **Clinical-grade performance**: Sub-second processing times
- **Intelligent measurement**: Size, area, and morphometric analysis with temporal stabilization
- **Adaptive visualization**: Real-time heatmaps that adjust to camera movement
- **Real-time inference**: GPU-accelerated deployment ready

---

## πŸŽ₯ Live System Demonstration

**Real-time EndoSight AI in action** - showcasing automated polyp detection, intelligent segmentation, stabilized measurement analysis, and clinical classification on endoscopy footage.
"""
    
    footer = """
---

## πŸš€ Development Status & Impact

βœ… **Model Training**: Complete with validation metrics achieved  
πŸ”„ **Clinical Integration**: Active collaboration with gastroenterology clinic  
πŸ“Š **Performance**: Production-ready accuracy and speed benchmarks  
🎯 **Target**: Characterization of neoplastic versus non-neoplastic lesions

### 🀝 Research Collaboration

**Alumnus of**: Universidad Central de Venezuela
**Research Focus**: AI-Assisted Endoscopy & Computer-Aided Diagnosis  

### πŸ“ž Professional Contact

**Technical Lead**: Daniel Cavadia  
**Email**: [dan.cavadia@gmail.com](mailto:dan.cavadia@gmail.com)  

---

βš•οΈ **Medical Disclaimer**: *Research prototype under clinical validation - not intended for diagnostic use.*
"""

    # Create simple interface using Gradio 3.34.0 syntax
    with gr.Blocks(title="EndoSight AI") as demo:
        gr.Markdown(f"# {title}")
        gr.Markdown(description)
        
        # Simple video display (no fancy features in older version)
        video = gr.Video(
            value="demo_video.mp4",
            label="EndoSight AI Real-time Detection & Segmentation with Intelligent Measurements"
        )
        
        gr.Markdown("""
### πŸ” What You're Seeing:
- **Yellow bounding boxes**: Real-time polyp detection with smooth tracking
- **Fluid heatmaps**: Dynamic segmentation visualization that adapts to movement
- **Bottom-left panels**: Stabilized measurement displays with clinical classifications
- **Size categories**: Diminutive (<5mm), Small (5-9mm), Large (β‰₯10mm) classifications
- **Risk assessment**: Low/Moderate/High risk indicators based on polyp size
- **Measurement accuracy**: Diameter and area calculations with error margins
- **Anti-fluctuation**: Intelligent system prevents measurement swings during camera movement
- **Processing metrics**: Real-time FPS and confidence indicators
- **Multi-polyp detection**: Simultaneous analysis capability with unique tracking IDs
""")
        
        gr.Markdown(footer)
    
    return demo

if __name__ == "__main__":
    demo = create_demo()
    demo.launch()