BEN2 Introduction and Core Functions
BEN2 (Background Erase Network 2) is a deep learning model developed by Prama LLC dedicated to automatic image background removal. Its core innovation is the use ofConfidence Guided Matting (CGM) Pipeline, which achieves sub-pixel level foreground segmentation accuracy by the base model working in concert with the refinement network (processing low confidence pixels).
- List of core functions::
- Automatic generation of binary masks and de-contextualized foreground images
- High-precision processing of complex scenes such as hair/edges
- 4K UHD image support (20 sec/picture)
- CUDA GPU acceleration (1080p images in 6 seconds)
- Provide Python API for easy system integration
Notably, the model is trained on the DIS5k benchmark and 22K proprietary datasets, and its lightweight volume of 1.13GB gives it both high performance and ease of deployment.
This answer comes from the articleBEN2: Deep learning model for fast background removal from images, videosThe































