Optimal binary sequences for coded exposure photography using genetic algorithm
Kuihua Huang, Jinxin Hou, Jun Zhang · 2014
This paper proposes to use genetic algorithm (GA) to find the optimal binary sequences for coded exposure photography. The main improvement of the proposed method over previous methods mainly stems from using the optimally rotated and appended Legendre sequences as the initial population of GA. Jeon et al. creatively apply the modified Legendre sequence which is originally used in information theory community to coded exposure photography and improve the previous random search methods both in computational time and image deblurring quality. However, the flipping operation which tries to improve the sequence quality is not cost effective. We use the crossover and mutation operators in GA to substitute the flipping operation and obtain much gain in terms of both merit factor and coded factor. Experimental results with synthetic and real data demonstrate that coded exposure using sequences generated by our proposed method obtain better motion deblurring results compared to the state of the art methods.