An official website of the United States government
Recently, transformers have demonstrated great potential for modeling long-term dependencies from skeleton sequences and thereby gained ever-increasing attention in skeleton action recognition. However, the existing transformer-based approaches heavily rely on the naive attention mechanism for capturing the spatiotemporal features, which falls short in learning discriminative representations that exhibit similar motion patterns. To address this challenge, we introduce the Frequency-aware Mixed Transformer (FreqMixFormer), specifically designed for recognizing similar skeletal actions with subtle discriminative motions. First, we introduce a frequency-aware attention module to unweave skeleton frequency representations by embedding joint features into frequency attention maps, aiming to distinguish the discriminative movements based on their frequency coefficients. Subsequently, we develop a mixed transformer architecture to incorporate spatial features with frequency features to model the comprehensive frequency-spatial patterns. Additionally, a temporal transformer is proposed to extract the global correlations across frames. Extensive experiments show that FreqMiXFormer outperforms SOTA on 3 popular skeleton action recognition datasets, including NTU RGB+D, NTU RGB+D120, and NW-UCLA datasets. Our project is publicly available at: https://github.com/wenhanwu95/FreqMixFormer.
more »
« less
This content will become publicly available on January 30, 2026
FreqMixFormerV2: Lightweight Frequency-aware Mixed Transformer for Human Skeleton Action Recognition
Transformer-based human skeleton action recognition has been developed for years. However, the complexity and high parameter count demands of these models hinder their practical applications, especially in resource-constrained environments. In this work, we propose FreqMixForemrV2, which was built upon the Frequency-aware Mixed Transformer (FreqMixFormer) for identifying subtle and discriminative actions with pioneered frequency-domain analysis. We design a lightweight architecture that maintains robust performance while significantly reducing the model complexity. This is achieved through a redesigned frequency operator that optimizes high-frequency and low-frequency parameter adjustments, and a simplified frequency-aware attention module. These improvements result in a substantial reduction in model parameters, enabling efficient deployment with only a minimal sacrifice in accuracy. Comprehensive evaluations of standard datasets (NTU RGB+D, NTU RGB+D 120, and NW-UCLA datasets) demonstrate that the proposed model achieves a superior balance between efficiency and accuracy, outperforming state-of-the-art methods with only 60% of the parameters.
more »
« less
- Award ID(s):
- 1840080
- PAR ID:
- 10565647
- Publisher / Repository:
- IEEE International Conference on Automatic Face and Gesture Recognition
- Date Published:
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Self-supervised skeleton-based action recognition has attracted more attention in recent years. By utilizing the unlabeled data, more generalizable features can be learned to alleviate the overfitting problem and reduce the demand for massive labeled training data. Inspired by the MAE [1], we propose a spatial-temporal masked autoencoder framework for self-supervised 3D skeleton-based action recognition (SkeletonMAE). Following MAE's masking and reconstruction pipeline, we utilize a skeleton-based encoder-decoder transformer architecture to reconstruct the masked skeleton sequences. A novel masking strategy, named Spatial-Temporal Masking, is introduced in terms of both joint-level and frame-level for the skeleton sequence. This pre-training strategy makes the encoder output generalizable skeleton features with spatial and temporal dependencies. Given the unmasked skeleton sequence, the encoder is fine-tuned for the action recognition task. Extensive ex- periments show that our SkeletonMAE achieves remarkable performance and outperforms the state-of-the-art methods on both NTU RGB+D 60 and NTU RGB+D 120 datasets.more » « less
-
Existing multimodal-based human action recognition approaches are computationally intensive, limiting their deployment in real-time applications. In this work, we present a novel and efficient pose-driven attention-guided multimodal network (EPAM-Net) for action recognition in videos. Specifically, we propose eXpand temporal Shift (X-ShiftNet) convolutional architectures for RGB and pose streams to capture spatio-temporal features from RGB videos and their skeleton sequences. The X-ShiftNet tackles the high computational cost of the 3D CNNs by integrating the Temporal Shift Module (TSM) into an efficient 2D CNN, enabling efficient spatiotemporal learning. Then skeleton features are utilized to guide the visual network stream, focusing on keyframes and their salient spatial regions using the proposed spatial–temporal attention block. Finally, the predictions of the two streams are fused for final classification. The experimental results show that our method, with a significant reduction in floating-point operations (FLOPs), outperforms and competes with the state-of-the-art methods on NTU RGB-D 60, NTU RGB-D 120, PKU-MMD, and Toyota SmartHome datasets. The proposed EPAM-Net provides up to a 72.8x reduction in FLOPs and up to a 48.6x reduction in the number of network parameters. The code will be available at https://github.com/ahmed-nady/Multimodal-ActionRecognition.more » « less
-
Human skeleton-based action recognition offers a valuable means to understand the intricacies of human behavior because it can handle the complex relationships between physical constraints and intention. Although several studies have focused on encoding a skeleton, less attention has been paid to embed this information into the latent representations of human action. InfoGCN proposes a learning framework for action recognition combining a novel learning objective and an encoding method. First, we design an information bottleneck-based learning objective to guide the model to learn informative but compact latent representations. To provide discriminative information for classifying action, we introduce attention-based graph convolution that captures the context-dependent intrinsic topology of human action. In addition, we present a multi-modal representation of the skeleton using the relative position of joints, designed to provide complementary spatial information for joints. InfoGcn 1 1 Code is available at github.com/stnoahl/infogcn surpasses the known state-of-the-art on multiple skeleton-based action recognition benchmarks with the accuracy of 93.0% on NTU RGB+D 60 cross-subject split, 89.8% on NTU RGB+D 120 cross-subject split, and 97.0% on NW-UCLA.more » « less
-
Recent advances in Neural Radiance Field (NeRF)-based methods have enabled high-fidelity novel view synthesis for video with dynamic elements. However, these methods often require expensive hardware, take days to process a second-long video and do not scale well to longer videos. We create an end-to-end pipeline for creating dynamic 3D video from a monocular video that can be run on consumer hardware in minutes per second of footage, not days. Our pipeline handles the estimation of the camera parameters, depth maps, 3D reconstruction of dynamic foreground and static background elements, and the rendering of the 3D video on a computer or VR headset. We use a state-of-the-art visual transformer model to estimate depth maps which we use to scale COLMAP poses and enable RGB-D fusion with estimated depth data. In our preliminary experiments, we rendered the output in a VR headset and visually compared the method against ground-truth datasets and state-of-the-art NeRF-based methods.more » « less
