ABCD : Arbitrary Bitwise Coefficient for De-Quantization

Woo Kyoung Han, Byeonghun Lee, Sanghyun Park, Kyong Hwan Jin
2/14/2026

No official code found. Best implementation: WooKyoungHan/ABCD (29 stars, BSD-3-Clause, PyTorch, last commit 2023-06-01).

Abstract

Modern displays and contents support more than 8bits image and video. However, bit-starving situations such as compression codecs make low bit-depth (LBD) images (<8bits), occurring banding and blurry artifacts. Previous bit depth expansion (BDE) methods still produce unsatisfactory high bit-depth (HBD) images. To this end, we propose an implicit neural function with a bit query to recover de-quantized images from arbitrarily quantized inputs. We develop a phasor estimator to exploit the information of the nearest pixels. Our method shows superior performance against prior BDE methods on natural and animation images. We also demonstrate our model on YouTube UGC datasets for de-banding. Our source code is available at https://github.com/WooKyoungHan/ABCD

DOISemantic Scholar

Code Implementations(1)

WooKyoungHan/ABCDCommunity60%
PyTorchBSD-3-Clause

Official implementation of the CVPR'23 Paper "ABCD : Arbitrary Bitwise Coefficient for De-quantization"

294PythonMar 7, 20233 years agoBSD-3-Clause
cvpr2023deep-learningdequantizationimage-processingimage-restoration+2 more

Cite this paper

@article{han2026abcd,
  title  = {ABCD : Arbitrary Bitwise Coefficient for De-Quantization},
  author = {Woo Kyoung Han and Byeonghun Lee and Sanghyun Park and Kyong Hwan Jin},
  year   = {2026},
  doi    = {10.1109/CVPR52729.2023.00569},
  url    = {https://doi.org/10.1109/CVPR52729.2023.00569},
  journal = {CVPR 2023 2023}
}

Discussion