Accelerating feature detectors for real-time vision-based applications
Khine, Thaw Hnin
Date of Issue2016
School of Computer Engineering
In computer vision system, the features detection and extraction is one of the most basic and important step in performing real-time applications such as object recognition and motion tracking. Among the feature detection methods, Harris corner detection is one of the widely use algorithm as an early processing step. There are several implementation of Harris corner detection in different software platform. However this software implementation requires long computation time because of the usage of multiple repetitive computations. In addition, software implementation is probably not compatible with real-time low cost processor. Therefore, this paper purposes an efficient hardware approach that offloads the repetitive feature detection procedures into logic gates. Hence the solution is low cost to produce and less complexity to operate compared to its software counterpart. The experiments and demostrations in this project show that the speed and accuracy of the accelerated feature detector are good enough for many real world applications.
Final Year Project (FYP)
Nanyang Technological University