Pickpocketing Recognition in Still Images
Human activity recognition (HAR) is a challenging topic in the computer vision eld. Pickpocketing is a type of human criminal ac- tions. It needs extensive research and development for detection. This paper investigates the possibility of pickpocketing recognition in still im- ages. This paper takes into consideration classi cation and detection. We develop our models from state-of-art pre-trained models: VGG16, ResNet50, ResNet101, and ResNet152. Moreover, we also include a con- volutional block attention module (CBAM ) in the model. The atten- tion mechanism enhances model performances by focusing on informative features. For classi cation, the highest accuracy (89%) is ResNet152 with CBAM  (ResNet152+CBAM). We also examine pickpocketing detec- tion on RetinaNet  and YOLOv.3 . The mean average precision (mAP) of pickpocketing detection is consistent with Redmon et al. . RetinaNet's precision (80 mAP) is higher than YOLOv.3 (78 mAP), but YOLOv.3 is much faster detection. ResNet152 + CBAM model detection on RetinaNet approach provides the highest mAP. However, it is much slower detection than YOLOv.3 (only 10ms). This paper proves that It is possible to implement pickpocketing on still images in a reliable time and with outstanding accuracy. This proposed model can be applied to the other HAR tasks.
Citation : Damrongsiri P., Malekmohamadi H. (2021) Pickpocketing Recognition in Still Images. In: Del Bimbo A. et al. (eds) Pattern Recognition. ICPR International Workshops and Challenges. ICPR 2021. Lecture Notes in Computer Science, vol 12664. Springer, Cham. https://doi.org/10.1007/978-3-030-68799-1_11
Research Institute : Institute of Artificial Intelligence (IAI)
Peer Reviewed : Yes