This work presents a xed point simulation of Compressed Sensing (CS) and reconstruction for Super-Resolutiontask using Image System Engineering Toolbox (ISET). This work shows that performance of CS for super-resolution in xed point implementation is similar to oating point implementation and there is negligible loss inreconstruction quality. It also shows that CS Super-Resolution requires much less computation eort comparedto CS using Gaussian Random matrices. Additionally, it also studies the eect of Analog-to-Digital-Converter(ADC) bitwidth and image sensor noise on reconstruction performance. CS super-resolution cuts the raw databits generated from image sensor by more than half and conversion of reconstruction algorithm to xed pointallows one to simplify the hardware implementation by replacing expensive oating point computational unitswith faster and energy ecient xed point units.
展开▼