Tags:
Dr. Hamdan Awan, postdoctoral research fellow in TSSG, has co-authored a paper outlining the significance of visual object tracking with many societal and industry benefits such as improved home security and a reduction in human errors making an assembly line more effective.
Visual object tracking aims to track the objects in video sequences as they move across the frames of video. These object may be people, cars, animals or any thing that appears in the video. This
becomes very important nowadays where example applications may include surveillance to detect suspicious activity, video analysis to extract the highlights in sports (e.g., tracking a favourite player in a football match), and last but not least human computer interface to help the visually impaired. Tracking an object in a complex environment is a very difficult challenge. A complex environment is determined as having a lot of occlusion i.e. deformation in an object and multiple objects of the same features.
To increase the robustness of the visual object tracking algorithm by handling the occlusion effectively, we first propose a strategy to detect the occlusion using two cues from the response map i.e. peak correlation score and peak to side lobe ratio. After successful detection of tracking failure, the second strategy is proposed to save the target from resulting in higher errors. Our algorithm shows significant improvement in the tracking accuracy over videos, specifically focusing on addressing six challenges; background clutters, occlusion, deformation, scale variation, out of plane rotation and motion blur.
Nowadays it’s common to use CCTV cameras for home security but it takes a lot of memory if you are recording and saving the videos 24/7. So, object tracking can help to record and save only suspicious activity when it happens. Similarly, you can track the activities of kids when you are not at home. In other words, this research allows the computer to obtain a better model of the real world.
Object tracking has been around for years, but is becoming more apparent across a range of industries now more than ever before. In the manufacturing industry, manual sorting involves high cost of labour and accompanying human errors. Even with robots, the process is not accurate enough and is still prone to discrepancies. With AI-powered object tracking, the objects are classified as per the parameter selected by the manufacturer and statistics of the number of objects displayed. It significantly reduces the abnormalities in categorization and makes the assembly line more flexible.
Though a lot of researchers are working in the visual object tracking field, the area still requires a considerable amount of attention to attain 100 % efficiency.
Publication Title: AFAM-PEC: Adaptive Failure Avoidance Tracking Mechanism using Prediction-Estimation Collaboration
Authors: Hamdan Awan, Baber Khan, Ahmad Ali, Abdul Jalil, Khizer Mehmood and Maria Murad
Publication Date: 10 August 2020
Journal: IEEE Access
Link to publication: https://ieeexplore.ieee.org/abstract/document/9163354]]>