Research map / Video / Action recognition
Adaptive Residual Quantization for Memory-Efficient Temporal Action Segmentation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes adaptive residual quantization to compress stored pre-extracted frame features for memory-efficient temporal action segmentation, especially for replay in incremental TAS.
Problem
Storing and replaying pre-extracted frame features for temporal action segmentation becomes a memory bottleneck for long-form video, particularly in incremental TAS where past-task data must be retained to mitigate catastrophic forgetting.
Approach
Uses adaptive residual quantization of frame features to reduce the storage footprint of replay banks, as an alternative to dense frame replay and condensation methods (the abstract is truncated, so the precise mechanism is not stated).
Closest papers by meaning
- Condensing Action Segmentation Datasets via Generative Network Inversion (CVPR 2025 · similarity 0.87)
- Hierarchical Vector Quantization for Unsupervised Action Segmentation (AAAI 2025 · similarity 0.87)
- Efficient and Effective Weakly-Supervised Action Segmentation via Action-Transition-Aware Boundary Alignment (CVPR 2024 · similarity 0.87)
- TTVidT: Decoupling the Temporal Axis for Efficient Motion-Centric Video Pretraining (NeurIPS 2026 · similarity 0.87)
- End-to-End Temporal Action Detection with 1B Parameters Across 1000 Frames (CVPR 2024 · similarity 0.87)
- FACT: Frame-Action Cross-Attention Temporal Modeling for Efficient Action Segmentation (CVPR 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.