Nature-Inspired BASIS Feature Descriptor and Its Hardware Implementation

This paper presents the development of a color feature descriptor wellsuited for low resource applications such as UAV embedded systems, small microprocessors, and field programmable gate array (FPGA) fabric, and its hardware implementation. The BAsis Sparse-coding Inspired Similarity (BASIS) descriptor utilizes sparse coding to create dictionary images that model the regions in the human visual cortex. Due to the reduced amount of computation required for computing BASIS descriptors, reduced descriptor size, and the ability to create the descriptors without the use of floating point, this approach is an excellent candidate for FPGA hardware implementation. The BASIS descriptor was tested on a dataset of real aerial images with the task of calculating a frame-to-frame homography. Experimental results show that the software and bit-level accurate BASIS descriptor outperforms SIFT and SURF. The bit-level accurate hardware version of the BASIS descriptor also results in no loss of accuracy when compared with the software version. BASIS descriptors are more space efficient than other descriptors, and can be computed entirely in FPGA fabric, allowing the descriptor to operate at real-time frame rates on a low power, embedded platform such as an FPGA.