Enhancing embedded AI-based object detection using multi-view approach

Zijie Ning (1, 2) , Mostafa Rizk (1) , Amer Baghdadi (1, 2) , Jean-Philippe Diguet (3)

1 Lab-STICC_2AI - Equipe Algorithm Architecture Interactions
2 IMT Atlantique - MEE - Département Mathematical and Electrical Engineering
3 CROSSING - FrenCh austRalian labOratory for humanS/autonomouS agents teamING

Object detection based on convolutional neural network (CNN) is widely used in multitude emergent applications. Yet, the deployment of CNNs on embedded devices at the edge with reduced resources and power budget poses a real challenge. In this paper, we address this issue by enhancing the detection performance without impacting the inference speed. We investigate the use of multi-view for the same scene to achieve better detection performance. A novel system of distributed smart cameras is proposed where each camera integrates a CNN for detection. Implementation results show that using light networks on the distributed cameras can lead to better detection performance and a reduction in the overall consumed power.

Mots clés: Multi-view; Object detection; Embedded edge device

Domaines: Informatique [cs]; Intelligence artificielle [cs.AI]; Calcul parallèle, distribué et partagé [cs.DC]; Traitement du signal et de l'image [eess.SP]; Electronique

Zijie Ning, Mostafa Rizk, Amer Baghdadi, Jean-Philippe Diguet. Enhancing embedded AI-based object detection using multi-view approach. RSP 2022: IEEE International Workshop on Rapid System Prototyping, part of Embedded Systems Week (ESWEEK), Oct 2022, Shanghai, China. ⟨10.1109/RSP57251.2022.10039026⟩. ⟨hal-03836472⟩