A Reversible Grayscale Method Based on Bit-Field Multi-Channel Fusion Encoding

This work proposes a reversible grayscale-compatible representation based on bit-field multi-channel fusion. The high 8 bits store a standard grayscale image so that the file remains directly viewable as grayscale, while the low 8 bits store compact hue and saturation hints for color restoration.

This work appears in 2025 10th International Conference on Image, Vision and Computing (ICIVC), 555-560.

Problem setting

This work proposes a reversible grayscale-compatible representation based on bit-field multi-channel fusion. The high 8 bits store a standard grayscale image so that the file remains directly viewable as grayscale, while the low 8 bits store quantized hue and saturation hints. A lightweight restoration network uses the grayscale carrier and color hints to reconstruct RGB, balancing grayscale compatibility, storage simplicity, and color recovery quality.

The figures below show the bit-field encoding design, the restoration network, and qualitative reconstruction examples for the reversible grayscale representation.

Method and visual evidence

The method treats grayscale compatibility as a representation problem. The high byte stores a conventional grayscale image for direct viewing, while the low byte stores compact color hints that guide RGB reconstruction.

Method overview.

Representation and setup.

Experimental evidence.

Result comparison.

Additional visual result.

Results and impact

The reported evaluation compares restored color images against the original RGB inputs while preserving standard grayscale display behavior. The code link points to the public implementation repository for the encoding and restoration workflow.