What question did this study set out to answer?

This research aims to develop a novel convolutional neural network for real-time classification of surveillance videos, focusing on effective detection of anomalies.

June 19, 2026Open Access

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

KMK. MohanaranganNational Institute of Technology Tiruchirappalli PPP. PalanisamyNational Institute of Technology Tiruchirappalli

Key Points

This research aims to develop a novel convolutional neural network for real-time classification of surveillance videos, focusing on effective detection of anomalies.
Developed TriPath3DNet, a 3D CNN architecture integrating three temporal pathways.
Evaluated performance on four datasets, using a lightweight ResNet3D-18 backbone.
Conducted ablation studies and visualized attention maps using Grad-CAM.
Achieved up to 95.37% accuracy and 99.42% AUC with an inference latency of 129 to 137 ms per 50-frame clip.
Outperformed baseline models, especially in detecting anomalies in UCF-Crime.
Demonstrated superior calibration for anomaly detection with significant performance margins.

Abstract

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.

Ask AI

Helpful

Bookmark

View Full Paper