Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Sun, Zepeng, Zheng, Naichuan, Xia, Hailun, Wu, Junjie, Bao, Liwei, Zhang, Xiaotai
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2604.18274
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866913063883505664
author Sun, Zepeng
Zheng, Naichuan
Xia, Hailun
Wu, Junjie
Bao, Liwei
Zhang, Xiaotai
author_facet Sun, Zepeng
Zheng, Naichuan
Xia, Hailun
Wu, Junjie
Bao, Liwei
Zhang, Xiaotai
contents Temporal Action Detection (TAD) requires precise localization of action boundaries within long, untrimmed video sequences. While current high-performing methods achieve strong accuracy, they are often characterized by excessive parameter counts, substantial computational overhead, and a reliance on specialized operators that hinder deployment across diverse hardware platforms. This paper presents LiquidTAD, a framework that distills the exponential relaxation prior of liquid neural dynamics into a parallel temporal operator, rather than reproducing full Liquid Neural Network (LNN) dynamics. By introducing a Parallel Liquid-inspired Relaxation mechanism, sequential ODE solving is avoided through a fully vectorized, non-recursive formulation built entirely upon standard neural operations, enabling hardware-agnostic deployment with linear complexity with respect to the temporal length. A complementary Hierarchical Decay-Rate Sharing Strategy further adapts this relaxation prior across feature pyramid levels, stabilizing optimization and implicitly compensating for temporal compression in deeper layers. Experimental evaluations on THUMOS-14 and ActivityNet-1.3 demonstrate that LiquidTAD achieves accuracy competitive with strong baselines while substantially lowering the model footprint. Specifically, on THUMOS-14, LiquidTAD achieves 69.46\% average mAP with only 10.82M parameters and 27.17G FLOPs, reducing the parameter count by over 60\% compared with ActionFormer.
format Preprint
id arxiv_https___arxiv_org_abs_2604_18274
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle LiquidTAD: Efficient Temporal Action Detection via Parallel Liquid-Inspired Temporal Relaxation
Sun, Zepeng
Zheng, Naichuan
Xia, Hailun
Wu, Junjie
Bao, Liwei
Zhang, Xiaotai
Computer Vision and Pattern Recognition
Temporal Action Detection (TAD) requires precise localization of action boundaries within long, untrimmed video sequences. While current high-performing methods achieve strong accuracy, they are often characterized by excessive parameter counts, substantial computational overhead, and a reliance on specialized operators that hinder deployment across diverse hardware platforms. This paper presents LiquidTAD, a framework that distills the exponential relaxation prior of liquid neural dynamics into a parallel temporal operator, rather than reproducing full Liquid Neural Network (LNN) dynamics. By introducing a Parallel Liquid-inspired Relaxation mechanism, sequential ODE solving is avoided through a fully vectorized, non-recursive formulation built entirely upon standard neural operations, enabling hardware-agnostic deployment with linear complexity with respect to the temporal length. A complementary Hierarchical Decay-Rate Sharing Strategy further adapts this relaxation prior across feature pyramid levels, stabilizing optimization and implicitly compensating for temporal compression in deeper layers. Experimental evaluations on THUMOS-14 and ActivityNet-1.3 demonstrate that LiquidTAD achieves accuracy competitive with strong baselines while substantially lowering the model footprint. Specifically, on THUMOS-14, LiquidTAD achieves 69.46\% average mAP with only 10.82M parameters and 27.17G FLOPs, reducing the parameter count by over 60\% compared with ActionFormer.
title LiquidTAD: Efficient Temporal Action Detection via Parallel Liquid-Inspired Temporal Relaxation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2604.18274