Variable-Rate Deep Image Compression based on Low-Rank Adaptation by Progressive Learning
2026-06-15 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionMultimedia
AI summaryⓘ
The authors focus on improving how images are compressed at different quality levels using less memory and training effort. They propose adding a special module called LoRAM to existing deep image compression methods, which adapts the compression rate efficiently without making the system slower. Their approach uses a technique called Low-Rank Adaptation to fine-tune models progressively, saving lots of storage and training data compared to using multiple separate models. Experiments show their method performs similarly to traditional ones but with much less resource use.
image compressionvariable-rate compressiondeep image compressionLow-Rank Adaptation (LoRA)LoRA Rate-Adaptive Module (LoRAM)parameter-efficient fine-tuningprogressive learninginference complexity
Authors
Xing-Yu Xu, Chen-Hsiu Huang, Ja-Ling Wu
Abstract
In the digital age, image compression is crucial for numerous applications, including web media, streaming services, high-resolution medical imaging, and connected vehicle networks, enabling efficient data storage and transmission. With the increasing demand for high-quality image communication, the need for advanced compression techniques becomes increasingly critical. Numerous Deep Image Compression (DIC) techniques have recently been introduced, showing impressive performance compared to traditional standards. However, variable-rate image compression remains an unresolved issue. Specific DIC methods deploy multiple networks to attain different compression rates, whereas others use a single model, which often results in higher computational complexity and reduced performance. This work proposes a progressive learning approach for variable-rate image compression based on the parameter-efficient fine-tuning method, the Low-Rank Adaptation (LoRA). We introduce an additional LoRA Rate-Adaptive Module (LoRAM) in DIC methods. Due to the re-parameterized merging of LoRA, our proposed method does not introduce additional computational complexity during inference. Compared to methods utilizing multiple models, comprehensive experiments demonstrate that our approach achieves competitive performance, saving 99\% in parameter storage, 90% in datasets, and 97% in training steps.