What Matters in Practical Learned Image Compression
2026-05-06 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionArtificial IntelligenceMachine Learning
AI summaryⓘ
The authors studied how to build a learned image compression system that balances how good the pictures look with how fast it runs on devices like phones. They tested many different design options and used automated search to find the best model configurations for both speed and image quality. Their final system compresses images much more efficiently than popular codecs and other learned methods while running very fast on devices such as the iPhone 17 Pro Max. User studies showed it saves a lot of data compared to existing standards without slowing down encoding or decoding.
learned codecsimage compressionperceptual qualityneural architecture searchbitrate savingsruntime optimizationAV1VVCsubjective user studiesML-based codecs
Authors
Kedar Tatwawadi, Parisa Rahimzadeh, Zhanghao Sun, Zhiqi Chen, Ziyun Yang, Sanjay Nair, Divija Hasteer, Oren Rippel
Abstract
One of the major differentiators unlocked by learned codecs relative to their hard-coded traditional counterparts is their ability to be optimized directly to appeal to the human visual system. Despite this potential, a perceptual yet practical image codec is yet to be proposed. In this work, we aim to close this gap. We conduct a comprehensive study of the key modeling choices that govern the design of a practical learned image codec, jointly optimized for perceptual quality and runtime -- including within the ablations several novel techniques. We then perform performance-aware neural architecture search over millions of backbone configurations to identify models that achieve the target on-device runtime while maximizing compression performance as captured by perceptual metrics. We combine the various optimizations to construct a new codec that achieves a significantly improved tradeoff between speed and perceptual quality. Based on rigorous subjective user studies, it provides 2.3-3x bitrate savings against AV1, AV2, VVC, ECM and JPEG-AI, and 20-40% bitrate savings against the best learned codec alternatives. At the same time, on an iPhone 17 Pro Max, it encodes 12MP images as fast as 230ms, and decodes them in 150ms -- faster than most top ML-based codecs run on a V100 GPU.