Ajayk/car-background-removal-birefnet

Car Background Removal BiRefNet

Fine-tuned BiRefNet checkpoint for segmenting cars and producing background masks.

Usage

Use the inference application at https://github.com/AjayK47/car-background-removal. It downloads model.safetensors automatically on first run.

Files

• model.safetensors: fine-tuned checkpoint, 844 MiB.
• config.json: training/export configuration retained with the checkpoint.

Intended Use

This model is intended for foreground segmentation of vehicle images. It returns a per-pixel foreground mask that can be used for background removal or a transparent PNG cutout.

Limitations

Output quality varies with image composition, occlusion, lighting, reflections, and vehicle type. It should not be used as the sole basis for safety-critical or high-stakes decisions.

Provenance And Licensing

This checkpoint was fine-tuned from ZhengPeng7/BiRefNet. Review the base-model license and the licenses of all fine-tuning data before redistributing or using this model commercially.

SHA-256 for model.safetensors:

d10aea1fbee3c059b90eaab3399ca5ed4121c0c3647796d588a0c9510c383c5b