Update README.md

README.md

@@ -32,25 +32,25 @@ Upload any road image and get instant visual annotations along with a structured
 
 ## 🎯 Damage Classes
 
-| Code | Class Name         | Color    |
-|------|--------------------|----------|
-| D00  | Longitudinal Crack | 🔵 Blue  |
-| D10  | Transverse Crack   | 🟢 Green |
-| D20  | Alligator Crack    | 🟠 Orange|
-| D40  | Pothole            | 🔴 Red   |
+| Code | Class Name | Color |
+|------|---------------------|------------|
+| D00 | Longitudinal Crack | 🔵 Blue |
+| D10 | Transverse Crack | 🟢 Green |
+| D20 | Alligator Crack | 🟠 Orange |
+| D40 | Pothole | 🔴 Red |
 
 ---
 
 ## 🌍 Supported Countries (Pothole Context)
 
-| Country            | Pothole Density          |
-|--------------------|--------------------------|
-| 🇮🇳 India          | ⬛⬛⬛⬛⬛ Very High      |
-| 🇳🇴 Norway         | ⬛⬛⬛⬛ High             |
-| 🇺🇸 United States  | ⬛⬛⬛ Moderate           |
-| 🇨🇿 Czech Republic | ⬛⬛⬛ Moderate           |
-| 🇨🇳 China          | ⬛⬛ Low-Moderate         |
-| 🇯🇵 Japan          | ⬛ Low                   |
+| Country | Pothole Density |
+|-------------------|------------------------|
+| 🇮🇳 India | ⬛⬛⬛ Very High |
+| 🇳🇴 Norway | ⬛⬛⬛ High |
+| 🇺🇸 United States | ⬛⬛⬛ Moderate |
+| 🇨🇿 Czech Republic | ⬛⬛⬛ Moderate |
+| 🇨🇳 China | ⬛⬛ Low-Moderate |
+| 🇯🇵 Japan | ⬛ Low |
 
 ---
 
@@ -69,40 +69,44 @@ Upload any road image and get instant visual annotations along with a structured
 ## 📊 Output Explained
 
 ### Detection Summary
+
 A per-class breakdown table showing:
+
 - Count of each damage type detected
 - Percentage of total detections
 
 ### Pothole (D40) Report
 
-| Field             | Description                                      |
-|-------------------|--------------------------------------------------|
-| Pothole Count     | Number of D40 boxes detected                     |
-| % of Detections   | Proportion of potholes vs all damage             |
-| Severity          | Critical / Moderate / Low / None                 |
-| Avg Confidence    | Mean confidence score for pothole boxes          |
+| Field | Description |
+|------------------|------------------------------------------------|
+| Pothole Count | Number of D40 boxes detected |
+| % of Detections | Proportion of potholes vs all damage |
+| Severity | Critical / Moderate / Low / None |
+| Avg Confidence | Mean confidence score for pothole boxes |
 
 **Severity thresholds:**
+
 - 🔴 **CRITICAL** — Potholes > 30% of all detections → Immediate repair needed
 - 🟠 **MODERATE** — Potholes > 10% → Schedule maintenance
 - 🟡 **LOW** — Any potholes detected → Monitor road surface
 - 🟢 **NONE DETECTED** — Road surface OK
 
 ### Country Context
+
 Provides pothole density, dominant damage types, and data collection method for the selected country as documented in the RDD2022 dataset.
 
 ---
 
 ## 🧠 Model Details
 
-| Property       | Value                    |
-|----------------|--------------------------|
-| Architecture   | YOLOv8s (small)          |
-| Weights file   | `yolov8s_best.pt`        |
-| Input size     | 640 × 640 px             |
-| Classes        | 4 (D00, D10, D20, D40)   |
-| Framework      | Ultralytics YOLOv8       |
-| Device         | CUDA (GPU) / CPU fallback|
+| Property | Value |
+|---------------|---------------------------|
+| Architecture | YOLOv8s (small) |
+| Weights file | `yolov8s_best.pt` |
+| Input size | 640 × 640 px |
+| Classes | 4 (D00, D10, D20, D40) |
+| Framework | Ultralytics YOLOv8 |
+| Device | CUDA (GPU) / CPU fallback |
 
 ---
 
@@ -116,22 +120,23 @@ Provides pothole density, dominant damage types, and data collection method for
 ---
 
 ## 🗂️ Repository Structure
+
+```text
 Automated-Road-Damage-Detection/
-├── app.py # Gradio UI + inference pipeline
-├── predict.py # Standalone prediction script
-├── yolov8s_best.pt # Fine-tuned YOLOv8s weights (22.5 MB)
-├── rdd2022.yaml # Dataset config (class names, paths)
-├── requirements.txt # Python dependencies
-├── examples/ # Sample road images (6 countries)
-│ ├── india_test_image.jpg
-│ ├── China_Drone_000253.jpg
-│ ├── United_States_004798.jpg
-│ ├── Czech_test_image.jpg
-│ ├── China_Drone_000295.jpg
-│ └── norway_road_test.jpg
+├── app.py                  # Gradio UI + inference pipeline
+├── predict.py              # Standalone prediction script
+├── yolov8s_best.pt         # Fine-tuned YOLOv8s weights (22.5 MB)
+├── rdd2022.yaml            # Dataset config (class names, paths)
+├── requirements.txt        # Python dependencies
+├── examples/               # Sample road images (6 countries)
+│   ├── india_test_image.jpg
+│   ├── China_Drone_000253.jpg
+│   ├── United_States_004798.jpg
+│   ├── Czech_test_image.jpg
+│   ├── China_Drone_000295.jpg
+│   └── norway_road_test.jpg
 └── README.md
-
-text
+```
 
 ---
 
@@ -154,6 +159,8 @@ The app will be available at `http://localhost:7860`.
 ---
 
 ## 📦 Requirements
+
+```text
 gradio>=4.0
 ultralytics
 torch
@@ -161,8 +168,7 @@ torchvision
 numpy
 Pillow
 opencv-python
-
-text
+```
 
 ---