See-through — anysplit's Auto Body Split is built on See-through, a framework for single-image layer decomposition of anime characters (github.com/shitagaki-lab/see-through).
Jian Lin, Chengze Li, Haoyun Qin, Kwun Wang Chan, Yanghua Jin, Hanyuan Liu, Stephen Chun Wang Choy, and Xueting Liu. 2026. See-through: Single-image Layer Decomposition for Anime Characters. In SIGGRAPH Conference Papers '26. ACM. https://doi.org/10.1145/3799902.3811209
The See-through code is licensed under the Apache License 2.0 and is provided on an "AS IS" basis, without warranties or conditions of any kind. anysplit is not affiliated with or endorsed by the See-through authors or their institutions.
See-through body-part model — Auto Body Split uses the body-part parsing model of See-through (24yearsold/l2d_sam_iter2), a learned segmentation model fine-tuned from SAM-HQ (github.com/SysCV/sam-hq). Its inference code is taken from the See-through repository, and parts of it come from Segment Anything by Meta. The code and the weights are licensed under the Apache License 2.0 and are provided on an "AS IS" basis, without warranties or conditions of any kind. anysplit is not affiliated with or endorsed by the See-through or SAM-HQ authors, Meta or their institutions.
SAM 2 — interactive selection (click and drag) uses SAM 2 by Meta (github.com/facebookresearch/sam2).
Nikhila Ravi, Valentin Gabeur, Yuan-Ting Hu, Ronghang Hu, Chaitanya Ryali, Tengyu Ma, Haitham Khedr, Roman Rädle, Chloe Rolland, Laura Gustafson, Eric Mintun, Junting Pan, Kalyan Vasudev Alwala, Nicolas Carion, Chao-Yuan Wu, Ross Girshick, Piotr Dollár, and Christoph Feichtenhofer. SAM 2: Segment Anything in Images and Videos. arXiv:2408.00714, 2024.
The SAM 2 code and model checkpoints are licensed under the Apache License, Version 2.0; its connected-components CUDA kernel is licensed under the BSD 3-Clause License. anysplit is not affiliated with or endorsed by Meta.
Open-source libraries — the web app uses the following libraries under their own licenses (MIT License):