Trellis Coded Quantization for JPEG Compression

JPEG (Joint Photographic Experts Group) compression is widely used for the efficient storage and transmission of digital images. However, improving compression performance while maintaining image quality remains a challenge. This paper presents a study on the application of Trellis Coded Quantization (TCQ) techniques which are used in many video coding such as VVC and H.264 to enhance JPEG compression. Trellis coded quantization(TCQ), a robust quantization technique, reduces errors and enhances compression efficiency. This research implements TCQ in different stages of the JPEG compression pipeline, including Discrete Cosine Transform (DCT) coefficient quantization and entropy coding. The proposed method achieves progressive coding by adaptively allocating bit budgets to different frequency bands during quantization, in which way we can get better quantization indexes for coding. The trellis quantization algorithm optimizes the quantization process by considering both the rate-distortion performance and quantization noise shaping. Experimental results demonstrate that the TCQ-based approach outperforms conventional JPEG compression by 27.3% to 34% in terms of compression efficiency and image quality preservation in presented datasets with a little more compression time cost. Furthermore, the study investigates the impact of different parameters and configurations on the performance of TCQ. It explores trade-offs between compression efficiency, visual quality, and computational complexity. The findings indicate that TCQ can significantly enhance the compression performance of JPEG while maintaining competitive image quality. The research also discusses the limitations and potential extensions of TCQ in JPEG compression. Future work may focus on exploring adaptive strategies for bit allocation and trellis quantization optimization, as well as investigating the integration of TCQ with other advanced compression techniques. In conclusion, this paper presents a comprehensive investigation of Trellis Coded Quantization for JPEG compression. The results demonstrate the effectiveness of TCQ in improving compression efficiency and image quality, thereby contributing to the advancement of JPEG compression techniques.