CodecSight: Leveraging Video Codec Signals for Efficient Streaming VLM Inference

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zou, Yulin, Chen, Yan, Chen, Wenyan, Park, JooYoung, Nitin, Shivaraman, Tao, Luo, Romero, Francisco, Ustiugov, Dmitrii
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914460950593536
author Zou, Yulin
Chen, Yan
Chen, Wenyan
Park, JooYoung
Nitin, Shivaraman
Tao, Luo
Romero, Francisco
Ustiugov, Dmitrii
author_facet Zou, Yulin
Chen, Yan
Chen, Wenyan
Park, JooYoung
Nitin, Shivaraman
Tao, Luo
Romero, Francisco
Ustiugov, Dmitrii
contents Video streaming analytics is a crucial workload for vision-language model serving, but the high cost of multimodal inference limits scalability. Prior systems reduce inference cost by exploiting temporal and spatial redundancy in video streams, but they target either the vision transformer (ViT) or the LLM with a limited view, leaving end-to-end opportunities untapped. Moreover, existing methods incur significant overhead to identify redundancy, either through offline profiling and training or costly online computation, making them ill-suited for dynamic real-time streams. We present CodecSight, a codec-guided streaming video analytics system, built on a key observation that video codecs already extract the temporal and spatial structure of each stream as a byproduct of compression. CodecSight treats this codec metadata as a low-cost runtime signal to unify optimization across video decoding, visual processing, and LLM prefilling, with transmission reduction as an inherent benefit of operating directly on compressed bitstreams. This drives codec-guided patch pruning before ViT encoding and selective key-value cache refresh during LLM prefilling, both of which are fully online and do not require offline training. Experiments show that CodecSight achieves an improvement in throughput of up to 3$\times$, and a reduction of up to 87% in GPU compute over state-of-the-art baselines, maintaining competitive accuracy with only 0$\sim$8% F1 drop.
format Preprint
id arxiv_https___arxiv_org_abs_2604_06036
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle CodecSight: Leveraging Video Codec Signals for Efficient Streaming VLM Inference
Zou, Yulin
Chen, Yan
Chen, Wenyan
Park, JooYoung
Nitin, Shivaraman
Tao, Luo
Romero, Francisco
Ustiugov, Dmitrii
Distributed, Parallel, and Cluster Computing
Computer Vision and Pattern Recognition
Machine Learning
Video streaming analytics is a crucial workload for vision-language model serving, but the high cost of multimodal inference limits scalability. Prior systems reduce inference cost by exploiting temporal and spatial redundancy in video streams, but they target either the vision transformer (ViT) or the LLM with a limited view, leaving end-to-end opportunities untapped. Moreover, existing methods incur significant overhead to identify redundancy, either through offline profiling and training or costly online computation, making them ill-suited for dynamic real-time streams. We present CodecSight, a codec-guided streaming video analytics system, built on a key observation that video codecs already extract the temporal and spatial structure of each stream as a byproduct of compression. CodecSight treats this codec metadata as a low-cost runtime signal to unify optimization across video decoding, visual processing, and LLM prefilling, with transmission reduction as an inherent benefit of operating directly on compressed bitstreams. This drives codec-guided patch pruning before ViT encoding and selective key-value cache refresh during LLM prefilling, both of which are fully online and do not require offline training. Experiments show that CodecSight achieves an improvement in throughput of up to 3$\times$, and a reduction of up to 87% in GPU compute over state-of-the-art baselines, maintaining competitive accuracy with only 0$\sim$8% F1 drop.
title CodecSight: Leveraging Video Codec Signals for Efficient Streaming VLM Inference
topic Distributed, Parallel, and Cluster Computing
Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2604.06036