Implementation of Scale Invariant Feature Transform detector on FPGA for low?power wearable devices for prostheses control

In this paper we describe an FPGA implementation of the Scale Invariant Feature Transform (SIFT) algorithm. The is required as its a lightweight device which makes it ideal for vision-guided hybrid neuro-prostheses utilised upper limbs replacement. SIFT point detection needed computation coordinates object-to-grasp in wearable multi-camera system. A modified algorithm proposed and into C/C++ language on Xilinx ZCU102 board. hardware/software solution compared to other hardware or implementations with baseline software detector OpenSIFT. optimised gives average precision 0.84 recall 0.94 SIFT-point baseline. has lower dissipated power than solutions like CPU GPU, better computational speed. This allows processing medium-sized images real-time low consumption.