Scholarlink

<p>This paper presents <b>TriPath3DNet</b>, a novel, efficient, and interpretable 3D CNN architecture designed for real-time, multi-class classification of short, motion-triggered surveillance video clips under challenging real-world conditions–including occlusion, variable lighting, adverse weather, and subtle low-motion anomalies such as loitering or zone intrusion. TriPath3DNet integrates three complementary temporal pathways–short-term motion, long-term context, and Temporal Difference Encoding (TDE)–into a lightweight ResNet3D (R3D)-18 backbone to jointly model transient dynamics and sustained activities. Evaluated on four datasets–including the newly curated <b>Virat1-RC</b>, <b>Virat2-RC</b>, <b>UCF-Crime</b>, and our proprietary <b>In-House Dataset (IHD)</b>–TriPath3DNet achieves state-of-the-art or near state-of-the-art performance, with up to <b>95.37% accuracy</b>, <b>99.42% AUC</b>, and an inference latency of <b>129 to 137&#xa0;ms per 50-frame clip</b> (approx 2.6&#xa0;ms per frame) on an 11&#xa0;GB GPU, using only <b>33.46&#xa0;M parameters</b>. Notably, it outperforms both CNN- and transformer-based baselines–including MViTv1, MViTv2, and VideoSwin–by significant margins, especially on anomaly-dense benchmarks like UCF-Crime, where most vision transformers struggle. While MViTv2 achieves slightly higher accuracy on IHD (91.87% vs. 89.46%), TriPath3DNet delivers substantially better AUC (98.07% vs. 90.96%), indicating superior calibration for critical anomaly detection. Ablation studies confirm that each temporal branch contributes meaningfully to performance, and Grad-CAM visualizations demonstrate spatially precise and temporally coherent attention maps. By aligning architectural design with edge–cloud deployment constraints and the operational realities of industrial surveillance, our work bridges the gap between academic research and real-world video analytics.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

TriPath3DNet: an efficient real-time model for multi-class classification in real-life surveillance videos of fixed duration

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月17日

查看全文
  • K. Mohanarangan,
  • P. Palanisamy

摘要

This paper presents TriPath3DNet, a novel, efficient, and interpretable 3D CNN architecture designed for real-time, multi-class classification of short, motion-triggered surveillance video clips under challenging real-world conditions–including occlusion, variable lighting, adverse weather, and subtle low-motion anomalies such as loitering or zone intrusion. TriPath3DNet integrates three complementary temporal pathways–short-term motion, long-term context, and Temporal Difference Encoding (TDE)–into a lightweight ResNet3D (R3D)-18 backbone to jointly model transient dynamics and sustained activities. Evaluated on four datasets–including the newly curated Virat1-RC, Virat2-RC, UCF-Crime, and our proprietary In-House Dataset (IHD)–TriPath3DNet achieves state-of-the-art or near state-of-the-art performance, with up to 95.37% accuracy, 99.42% AUC, and an inference latency of 129 to 137 ms per 50-frame clip (approx 2.6 ms per frame) on an 11 GB GPU, using only 33.46 M parameters. Notably, it outperforms both CNN- and transformer-based baselines–including MViTv1, MViTv2, and VideoSwin–by significant margins, especially on anomaly-dense benchmarks like UCF-Crime, where most vision transformers struggle. While MViTv2 achieves slightly higher accuracy on IHD (91.87% vs. 89.46%), TriPath3DNet delivers substantially better AUC (98.07% vs. 90.96%), indicating superior calibration for critical anomaly detection. Ablation studies confirm that each temporal branch contributes meaningfully to performance, and Grad-CAM visualizations demonstrate spatially precise and temporally coherent attention maps. By aligning architectural design with edge–cloud deployment constraints and the operational realities of industrial surveillance, our work bridges the gap between academic research and real-world video analytics.