Fourier Compressor: Frequency-Domain Visual Token Compression for Vision-Language Models

arXiv:2508.06038v3 Announce Type: replace-cross Abstract: Vision-Language Models (VLMs) incur substantial computational overhead and inference latency due to the large number of vision tokens introduced by high-resolution image and video inputs. Existing parameter-free token compression methods typically rely on token selection or merging, yet they risk discarding substantial visual information or distorting the original representation distribution, resulting in pronounced performance degradatio