TrAction: Action Recognition with Sparse Trajectories

• URL: http://arxiv.org/abs/2606.03490v1
• Date: Tue, 02 Jun 2026 11:07:33 GMT
• Title: TrAction: Action Recognition with Sparse Trajectories
• Abstract summary: We investigate an efficient alternative input modality that is largely free of such biases by construction: sparse point trajectories.<n>We develop a simple transformer architecture for 2.5D trajectory-based recognition together with a masked-trajectory pretraining.<n>Our method reaches 45% top-1 on Something-Something V2 and 54% on EPIC-Kitchens-100, and surpasses V-JEPA on time-reversal sensitivity.
• Score: 43.37305832944302
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Modern action recognition models operate on memory- and compute-intensive dense RGB video volumes and frequently exploit appearance and background shortcuts, for example, predicting actions from objects or scenes instead of characteristic motion. We investigate an efficient alternative input modality that is largely free of such biases by construction: sparse point trajectories. To this end, we develop a simple transformer architecture for 2.5D trajectory-based recognition together with a masked-trajectory pretraining, which we show to substantially improve downstream action recognition accuracy. Despite using only a fraction of the dense RGB input, our method reaches 45% top-1 on Something-Something V2 and 54% on EPIC-Kitchens-100, and surpasses V-JEPA on time-reversal sensitivity. More importantly, we find trajectory features to be complementary to state-of-the-art appearance-based features. Fusing our pretrained model with DINOv2 and V-JEPA 2 improves top-1 accuracy on Something-Something V2 by 8.7 and 1.6 points, respectively. Code: https://github.com/ecker-lab/TrAction

Related papers

• Surformer v1: Transformer-Based Surface Classification Using Tactile and Vision Features [1.124958340749622]
  Surformer v1 is a transformer-based architecture designed for surface classification using structured tactile features and PCA-reduced visual embeddings extracted via ResNet-50.<n>The model integrates tactile-specific encoders with cross-modal attention layers, enabling rich interactions between vision and touch.<n>We trained both Surformer v1 and Multimodal CNN to examine the impact of feature-based versus image-based multimodal learning on classification accuracy and computational efficiency.
  arXiv  Detail & Related papers  (2025-08-07T00:59:33Z)
• An Efficient 3D Convolutional Neural Network with Channel-wise, Spatial-grouped, and Temporal Convolutions [3.798710743290466]
  We introduce a simple and very efficient 3D convolutional neural network for video action recognition.<n>We evaluate the performance and efficiency of our proposed network on several video action recognition datasets.
  arXiv  Detail & Related papers  (2025-03-02T08:47:06Z)
• DT-JRD: Deep Transformer based Just Recognizable Difference Prediction Model for Video Coding for Machines [48.07705666485972]
  Just Recognizable Difference (JRD) represents the minimum visual difference that is detectable by machine vision. We propose a Deep Transformer based JRD (DT-JRD) prediction model for Video Coding for Machines (VCM) The accurately predicted JRD can be used reduce the coding bit rate while maintaining the accuracy of machine tasks.
  arXiv  Detail & Related papers  (2024-11-14T09:34:36Z)
• TK-Planes: Tiered K-Planes with High Dimensional Feature Vectors for Dynamic UAV-based Scenes [58.180556221044235]
  We present a new approach to bridge the domain gap between synthetic and real-world data for unmanned aerial vehicle (UAV)-based perception. Our formulation is designed for dynamic scenes, consisting of small moving objects or human actions. We evaluate its performance on challenging datasets, including Okutama Action and UG2.
  arXiv  Detail & Related papers  (2024-05-04T21:55:33Z)
• DVANet: Disentangling View and Action Features for Multi-View Action Recognition [56.283944756315066]
  We present a novel approach to multi-view action recognition where we guide learned action representations to be separated from view-relevant information in a video. Our model and method of training significantly outperforms all other uni-modal models on four multi-view action recognition datasets.
  arXiv  Detail & Related papers  (2023-12-10T01:19:48Z)
• DOAD: Decoupled One Stage Action Detection Network [77.14883592642782]
  Localizing people and recognizing their actions from videos is a challenging task towards high-level video understanding. Existing methods are mostly two-stage based, with one stage for person bounding box generation and the other stage for action recognition. We present a decoupled one-stage network dubbed DOAD, to improve the efficiency for-temporal action detection.
  arXiv  Detail & Related papers  (2023-04-01T08:06:43Z)
• CNN-transformer mixed model for object detection [3.5897534810405403]
  In this paper, I propose a convolutional module with a transformer. It aims to improve the recognition accuracy of the model by fusing the detailed features extracted by CNN with the global features extracted by a transformer. After 100 rounds of training on the Pascal VOC dataset, the accuracy of the results reached 81%, which is 4.6 better than the faster RCNN[4] using resnet101[5] as the backbone.
  arXiv  Detail & Related papers  (2022-12-13T16:35:35Z)
• It Takes Two: Masked Appearance-Motion Modeling for Self-supervised Video Transformer Pre-training [76.69480467101143]
  Self-supervised video transformer pre-training has recently benefited from the mask-and-predict pipeline. We explicitly investigate motion cues in videos as extra prediction target and propose our Masked Appearance-Motion Modeling framework. Our method learns generalized video representations and achieves 82.3% on Kinects-400, 71.3% on Something-Something V2, 91.5% on UCF101, and 62.5% on HMDB51.
  arXiv  Detail & Related papers  (2022-10-11T08:05:18Z)
• Anchor-free Small-scale Multispectral Pedestrian Detection [88.7497134369344]
  We propose a method for effective and efficient multispectral fusion of the two modalities in an adapted single-stage anchor-free base architecture. We aim at learning pedestrian representations based on object center and scale rather than direct bounding box predictions. Results show our method's effectiveness in detecting small-scaled pedestrians.
  arXiv  Detail & Related papers  (2020-08-19T13:13:01Z)

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.