Collision aware method boosts image fitting with fewer parameters
CARA: Collision-Aware Resolution Adaptation for Multiresolution Hash Encoding Based Image Fitting
Computer Vision and Pattern RecognitionGraphics
Summary
Fitting images with neural networks can be inefficient because certain parts of the data get overloaded while others are underused. The authors describe a new way to adjust how these image details are stored by balancing the amount of information at each level, reducing conflicts in memory. They also introduce a technique to rearrange pixels to spread out data more evenly and prevent losses without needing bigger memory tables. Their method works well on very large images and achieves better quality using fewer memory resources.
What this means in practice
- •For computer vision engineers: Improve high-fidelity image reconstruction using fewer neural network parameters for large and complex images.
- •For medical imaging teams: Efficiently store and process extremely high-resolution whole-slide images by reducing hash collisions and improving parameter use.
Authors
Linfeng Ye, Zhixiang Chi, Shayan Mohajer Hamidi, En-hui Yang, Konstantinos N. Plataniotis
Abstract
Multiresolution hash encodings have recently enabled fast and high-fidelity implicit neural representations by storing multi-scale features in fixed-size hash tables along a geometric resolution schedule. However, the standard design is data-agnostic: different resolution levels receive identical hash-table capacity despite large differences in image frequency content. As a result, some levels experience severe hash collisions while others underutilize parameters, leading to inefficient capacity allocation. To address this issue, we propose Collision-Aware Resolution Adaptation (CARA), a method that assigns per-level resolutions by balancing the effective information load across hash levels. This adaptive allocation reduces capacity bottlenecks and improves parameter efficiency. In addition, we introduce an invertible pixel-shuffle transform that reduces hash load factors by redistributing spatial information, thereby mitigating collision-induced information loss without enlarging the hash tables. To support evaluation on extremely high-resolution data, we also curate, to the best of our knowledge, the first uncompressed whole-slide image dataset for academic research. Experiments on Kodak images, gigapixel natural images, and raw whole-slide images demonstrate that CARA consistently improves the fidelity-parameter trade-off. Our method matches state-of-the-art performance while using only $27.76%$ of the parameters, and achieves up to $6.11$ dB PSNR improvement at comparable parameter counts. Code is provided in the supplementary.