DVFL-Net: A Lightweight Distilled Video Focal Modulation Network for Spatio-Temporal Action Recognition

• URL: http://arxiv.org/abs/2507.12426v2
• Date: Fri, 18 Jul 2025 14:10:43 GMT
• Title: DVFL-Net: A Lightweight Distilled Video Focal Modulation Network for Spatio-Temporal Action Recognition
• Authors: Hayat Ullah, Muhammad Ali Shafique, Abbas Khan, Arslan Munir,
• Abstract summary: Video Modulation Network, DVFL-Net, distills knowledge from a large pre-trained teacher into a compact nano student model.<n>We evaluate DVFL-Net on UCF50, UCFMDB, H51, SSV2, and Kinetics benchmarking against state-of-the-art methods in Human Action Recognition (HAR)<n>The results confirm the superiority of DVFL-Net in achieving an optimal balance performance between memory and efficiency, demonstrating lower usage, reduced GFLOPs, and strong accuracy, making it a practical solution for real-time HAR applications.
• Score: 1.4785166764137887
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: The landscape of video recognition has evolved significantly, shifting from traditional Convolutional Neural Networks (CNNs) to Transformer-based architectures for improved accuracy. While 3D CNNs have been effective at capturing spatiotemporal dynamics, recent Transformer models leverage self-attention to model long-range spatial and temporal dependencies. Despite achieving state-of-the-art performance on major benchmarks, Transformers remain computationally expensive, particularly with dense video data. To address this, we propose a lightweight Video Focal Modulation Network, DVFL-Net, which distills spatiotemporal knowledge from a large pre-trained teacher into a compact nano student model, enabling efficient on-device deployment. DVFL-Net utilizes knowledge distillation and spatial-temporal feature modulation to significantly reduce computation while preserving high recognition performance. We employ forward Kullback-Leibler (KL) divergence alongside spatio-temporal focal modulation to effectively transfer both local and global context from the Video-FocalNet Base (teacher) to the proposed VFL-Net (student). We evaluate DVFL-Net on UCF50, UCF101, HMDB51, SSV2, and Kinetics-400, benchmarking it against recent state-of-the-art methods in Human Action Recognition (HAR). Additionally, we conduct a detailed ablation study analyzing the impact of forward KL divergence. The results confirm the superiority of DVFL-Net in achieving an optimal balance between performance and efficiency, demonstrating lower memory usage, reduced GFLOPs, and strong accuracy, making it a practical solution for real-time HAR applications.

Related papers

• TDS-CLIP: Temporal Difference Side Network for Efficient VideoAction Recognition [22.84073695186728]
  We propose a memory-efficient Temporal Difference Side Network ( TDS-CLIP) to balance knowledge transferring and temporal modeling.<n>Specifically, we introduce a Temporal Difference Adapter (TD-Adapter), which can effectively capture local temporal differences in motion features.<n>We also designed a Side Motion Enhancement Adapter (SME-Adapter) to guide the proposed side network in efficiently learning the rich motion information in videos.
  arXiv  Detail & Related papers  (2024-08-20T09:40:08Z)
• Promoting CNNs with Cross-Architecture Knowledge Distillation for Efficient Monocular Depth Estimation [4.242540533823568]
  Transformer models are usually computationally-expensive, and their effectiveness in light-weight models are limited compared to convolutions. We propose a cross-architecture knowledge distillation method for MDE, dubbed DisDepth, to enhance efficient CNN models with the supervision of state-of-the-art transformer models. Our method achieves significant improvements on various efficient backbones, showcasing its potential for efficient monocular depth estimation.
  arXiv  Detail & Related papers  (2024-04-25T07:55:47Z)
• Streaming Anchor Loss: Augmenting Supervision with Temporal Significance [5.7654216719335105]
  Streaming neural network models for fast frame-wise responses to various speech and sensory signals are widely adopted on resource-constrained platforms. We propose a new loss, Streaming Anchor Loss (SAL), to better utilize the given learning capacity by encouraging the model to learn more from essential frames.
  arXiv  Detail & Related papers  (2023-10-09T17:28:35Z)
• Video-FocalNets: Spatio-Temporal Focal Modulation for Video Action Recognition [112.66832145320434]
  Video-FocalNet is an effective and efficient architecture for video recognition that models both local global contexts. Video-FocalNet is based on a-temporal focal modulation architecture that reverses the interaction and aggregation steps of self-attention. We show that Video-FocalNets perform favorably well against state-of-the-art transformer-based models for video recognition on five large-scale datasets.
  arXiv  Detail & Related papers  (2023-07-13T17:59:33Z)
• Spatiotemporal Attention-based Semantic Compression for Real-time Video Recognition [117.98023585449808]
  We propose a temporal attention-based autoencoder (STAE) architecture to evaluate the importance of frames and pixels in each frame. We develop a lightweight decoder that leverages a 3D-2D CNN combined to reconstruct missing information. Experimental results show that ViT_STAE can compress the video dataset H51 by 104x with only 5% accuracy loss.
  arXiv  Detail & Related papers  (2023-05-22T07:47:27Z)
• Video Action Recognition Collaborative Learning with Dynamics via PSO-ConvNet Transformer [1.876462046907555]
  We propose a novel PSO-ConvNet model for learning actions in videos. Our experimental results on the UCF-101 dataset demonstrate substantial improvements of up to 9% in accuracy. Overall, our dynamic PSO-ConvNet model provides a promising direction for improving Human Action Recognition.
  arXiv  Detail & Related papers  (2023-02-17T23:39:34Z)
• STSM: Spatio-Temporal Shift Module for Efficient Action Recognition [4.096670184726871]
  We propose a plug-and-play Spatio-temporal Shift Module (STSM) that is both effective and high-performance. In particular, when the network is 2D CNNs, our STSM module allows the network to learn efficient Spatio-temporal features.
  arXiv  Detail & Related papers  (2021-12-05T09:40:49Z)
• Long-Short Temporal Contrastive Learning of Video Transformers [62.71874976426988]
  Self-supervised pretraining of video transformers on video-only datasets can lead to action recognition results on par or better than those obtained with supervised pretraining on large-scale image datasets. Our approach, named Long-Short Temporal Contrastive Learning, enables video transformers to learn an effective clip-level representation by predicting temporal context captured from a longer temporal extent.
  arXiv  Detail & Related papers  (2021-06-17T02:30:26Z)
• MVFNet: Multi-View Fusion Network for Efficient Video Recognition [79.92736306354576]
  We introduce a multi-view fusion (MVF) module to exploit video complexity using separable convolution for efficiency. MVFNet can be thought of as a generalized video modeling framework.
  arXiv  Detail & Related papers  (2020-12-13T06:34:18Z)
• Speaker Representation Learning using Global Context Guided Channel and Time-Frequency Transformations [67.18006078950337]
  We use the global context information to enhance important channels and recalibrate salient time-frequency locations. The proposed modules, together with a popular ResNet based model, are evaluated on the VoxCeleb1 dataset.
  arXiv  Detail & Related papers  (2020-09-02T01:07:29Z)
• A Real-time Action Representation with Temporal Encoding and Deep Compression [115.3739774920845]
  We propose a new real-time convolutional architecture, called Temporal Convolutional 3D Network (T-C3D), for action representation. T-C3D learns video action representations in a hierarchical multi-granularity manner while obtaining a high process speed. Our method achieves clear improvements on UCF101 action recognition benchmark against state-of-the-art real-time methods by 5.4% in terms of accuracy and 2 times faster in terms of inference speed with a less than 5MB storage model.
  arXiv  Detail & Related papers  (2020-06-17T06:30:43Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.