---
license: apache-2.0
language:
- en
base_model:
- Ultralytics/YOLO26
pipeline_tag: image-segmentation
---

# Car Damage Type Model

This repository contains the Car Damage Type Detection model. It uses a YOLO architecture to identify and classify specific types of physical damage on vehicles.

## Model Details

- **Model type:** Object Detection and Segmentation (YOLOv26 via Ultralytics)
- **Task:** Vehicle Damage Type Detection

## Intended Use

This model is intended for detecting and classifying physical damage on vehicles in images or video frames. It can be used as a standalone detector or integrated into a larger pipeline.

## Detected Classes

The model detects and localizes the following 6 classes of damage:
- `0`: **crack** - Cracks on car body panels or lights.
- `1`: **dent** - Indentations or structural dents.
- `2`: **glass shatter** - Shattered or completely broken glass (windshields, windows).
- `3`: **lamp broken** - Broken headlights or tail lights.
- `4`: **scratch** - Paint scratches or surface abrasions.
- `5`: **tire flat** - Deflated or flat tires.


![train_val_curves](https://cdn-uploads.huggingface.co/production/uploads/696b86293adc65f5022974c4/Ya2f96TpdDQi8ogv7fDp6.png)


![confusion_matrix_normalized](https://cdn-uploads.huggingface.co/production/uploads/696b86293adc65f5022974c4/imDvz3dAz-1pNQBOTGah0.png)