Crime scene classification from skeletal trajectory analysis in surveillance settings
This work addresses crime detection in surveillance footage, but it appears incremental as it builds on existing datasets and methods without demonstrating major breakthroughs.
The paper tackles human-related crime classification in surveillance videos by using skeletal joint trajectories as input, introducing methods to generate trajectory-level ground truth labels and a classification framework. Experiments show the feasibility of the task, though no concrete performance numbers are provided.
Video anomaly analysis is a core task actively pursued in the field of computer vision, with applications extending to real-world crime detection in surveillance footage. In this work, we address the task of human-related crime classification. In our proposed approach, the human body in video frames, represented as skeletal joints trajectories, is used as the main source of exploration. First, we introduce the significance of extending the ground truth labels for HR-Crime dataset and hence, propose a supervised and unsupervised methodology to generate trajectory-level ground truth labels. Next, given the availability of the trajectory-level ground truth, we introduce a trajectory-based crime classification framework. Ablation studies are conducted with various architectures and feature fusion strategies for the representation of the human trajectories. The conducted experiments demonstrate the feasibility of the task and pave the path for further research in the field.