Morphological image enhancement procedure design by using genetic programming

摘要: In this paper, we propose a genetic programming algorithm to design the morphological image enhancement procedure. Given group of operations and logical as function set, evolves produce rational procedure which can enhance input images. A novel mechanism combines ground truth method feature significance is brought forward evaluate performance images enhanced by generated procedures. each generation, best fitted individuals are selected on basis fitness values, some participate in crossover or mutation with probability. After outputs individual. Seven five used algorithm. Furthermore, structuring elements randomly varied whole pattern space. These methods promote expressive ability Examined binary extraction, more accurate intelligible than previous work. task gray scale enhancement, applied infrared finger vein region interest. More features extracted accuracy authentication promoted.