CVDec 2, 2025

Instant Video Models: Universal Adapters for Stabilizing Image-Based Networks

arXiv:2512.03014v1h-index: 4
Originality Incremental advance
AI Analysis

This work addresses temporal flickering and corruption issues in video processing for computer vision applications, representing an incremental improvement through a general adapter approach.

The authors tackled the problem of temporal inconsistency and corruption sensitivity in frame-based networks applied to video by introducing universal stability adapters and a training process that works with frozen base networks, achieving improved stability and robustness across multiple vision tasks.

When applied sequentially to video, frame-based networks often exhibit temporal inconsistency - for example, outputs that flicker between frames. This problem is amplified when the network inputs contain time-varying corruptions. In this work, we introduce a general approach for adapting frame-based models for stable and robust inference on video. We describe a class of stability adapters that can be inserted into virtually any architecture and a resource-efficient training process that can be performed with a frozen base network. We introduce a unified conceptual framework for describing temporal stability and corruption robustness, centered on a proposed accuracy-stability-robustness loss. By analyzing the theoretical properties of this loss, we identify the conditions where it produces well-behaved stabilizer training. Our experiments validate our approach on several vision tasks including denoising (NAFNet), image enhancement (HDRNet), monocular depth (Depth Anything v2), and semantic segmentation (DeepLabv3+). Our method improves temporal stability and robustness against a range of image corruptions (including compression artifacts, noise, and adverse weather), while preserving or improving the quality of predictions.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes