Built independently by an author, for readers. Read the story and support ChapterPal
ChapterPal Logo
Log inSign up

keyword

frequency-modulation feed-forward network

A frequency-modulation feed-forward network is a neural network component designed to dynamically adjust and balance different frequency characteristics within data representations. Unlike standard feed-forward networks that apply uniform transformations across all feature dimensions, this architecture selectively modulates distinct frequency sub-bands, such as high-frequency edge details and low-frequency structural regions. By adaptively scaling and transforming these components, often within frequency-aware vision transformer architectures, the network reduces latent representation redundancy and preserves critical textures and directional details to improve overall processing and reconstruction performance.

1 item

Frequency-Aware Transformer for Learned Image Compression

Frequency-Aware Transformer for Learned Image Compression

Han Li, Shaohui Li, Wenrui Dai, Chenglin Li, Junni Zou, Hongkai Xiong

OrganizationsShanghai Jiao Tong UniversityTsinghua University

Why you should read this

Introduces a frequency-aware transformer architecture for learned image compression that captures directional image details through multiscale frequency decomposition, outperforming the VTM-12.1 standard codec by more than 13% in BD-rate across benchmark datasets.

Learned image compression (LIC) has gained traction as an effective solution for image storage and transmission in recent years. However, existing LIC methods are redundant in latent representation due to limitations in capturing anisotropic frequency components and preserving directional details. To overcome these challenges, we propose a novel frequency-aware transformer (FAT) block that for the first time achieves multiscale directional ananlysis for LIC. The FAT block comprises frequency-decomposition window attention (FDWA) modules to capture multiscale and directional frequency components of natural images. Additionally, we introduce frequency-modulation feed-forward network (FMFFN) to adaptively modulate different frequency components, improving rate-distortion performance. Furthermore, we present a transformer-based channel-wise autoregressive (T-CA) model that effectively exploits channel dependencies. Experiments show that our method achieves state-of-the-art rate-distortion performance compared to existing LIC methods, and evidently outperforms latest standardized codec VTM-12.1 by 14.5%, 15.1%, 13.0% in BD-rate on the Kodak, Tecnick, and CLIC datasets.

Added

2026-09-26