AI and Robotics Ventures Co., Ltd. (ARV), a leading expert in artificial intelligence and robotics technology in Thailand, has achieved significant recognition with the acceptance of its research paper titled “Online Multi-Camera People Tracking with Spatial-Temporal Mechanism and Anchor-Feature Hierarchical Clustering” at the 2024 Computer Vision and Pattern Recognition Conference (CVPR). CVPR is the world’s premier conference in the field of computer vision, focusing on how computers can interpret images and videos, similar to the human visual system. ARV’s presentation at the conference was met with very positive feedback from attendees.
This research showcases, the development of a real-time multi-camera people-tracking system, designed to operate precisely across diverse environments, such as offices, factories, and homes. The underlying algorithms were rigorously evaluated by leading companies and research labs worldwide. ARV’s involvement in this competition offered a significant chance to evaluate the algorithms’ effectiveness, pinpointing both their strengths and areas needing improvement. The company intend to use this data to further refine the technology’s performance. This achievement underscores ARV’s expertise in artificial intelligence and robotics, reinforcing its commitment to advancing technologies that improve people’s quality of life.
Dr. Thana Slanvetpan,General Manager of AI and Robotics Ventures (ARV), explained that ARV’s Machine Learning team has been actively researching and developing a multi-camera people-tracking algorithm to integrate as a new feature into the company’s DeepZoom system. This advanced solution significantly enhances object tracking through CCTV, boosting its effectiveness and responsiveness in emergency situations. By leveraging spatial mechanisms and hierarchical clustering techniques, the system can accurately track individuals, even in low-light or crowded settings, thereby enhancing safety and surveillance for businesses, organizations, and the nation as a whole.
Dr. Thana Slanvetpan emphasized that multi-camera people-tracking technology is emerging as a critical tool for enhancing operational efficiency across various sectors. Beyond its role in safety and surveillance, this technology is driving operational improvements industry-wide. Key features of this technology include:
Integration of Advanced Algorithms: This encompasses several key technologies. Object Detection, where AI identifies objects such as people, vehicles, or structures in images or videos. Once detected, Object Tracking reduces processing resources by monitoring these objects. Human Pose Estimation further enhances the system by detecting and analyzing human movements in both images and videos. Re-identification technology recognizes individuals who have appeared in previous footage, while the Camera Association seamlessly connects multiple CCTV cameras to provide a continuous overview of individuals’ movements across different locations.
Utilization of Spatial and Temporal Data: This critical technique enhances accuracy and expands coverage, enabling the system to pinpoint precise locations and track movements in real-time. It can be deployed in a variety of settings, including office buildings, shopping malls, factories, and public spaces, offering complex and precise capabilities, particularly in real-time multi-camera cooperation. This functionality is invaluable for tracking objects or individuals, analyzing behavior, and detecting anomalies.
In addition, ARV is committed to supporting the Thai government’s initiative to position Thailand as a leading hub for technology and innovation. The company is focused on developing artificial intelligence and robotics technologies for practical applications across various sectors, including healthcare, agriculture, transportation, and energy. As part of this commitment, ARV continues to advance its multi-camera people tracking technology, while broadening its applications to deliver comprehensive solutions tailored to the specific needs of its customers.
For further inquiries or more information, please visit the website at https://arv.co.th/, the Facebook page at AI and Robotics Ventures (https://www.facebook.com/arv.th), or call 02 078 4000.
.jpg){kind=tracking}
