人臉識別外文翻譯---對于臉部識別系統(tǒng)研究遺傳算法-其他專業(yè)-全文預(yù)覽
【正文】 template matching method is also execution time shows the effectiveness of the GA based recognition method. Figures 2 and 3 are the original images and the templates for the values are the widthheight in pixels of the ,three images are presented,the content and size of which are 2a has two faces(the faces of a person and a toy), shows a face tipped to one side,and the person in wears a hat and the background is more plicated than in and b. The two templates in are not extracted from the same normal use,the template should be extracted as the average of several feature ,the template(a)0 is generated from(a)1,(a)2,and(a)3,and takes the average value of the gray levels from the three same is also true for(b)0. original images(max_xmax_y).a 238 185 275225 for matching(temp_x temp_y).a Template Template 2 3 Experiment and parison The geic operations and GA parameters are presented in Table 1 and Table fitness is defined as 255)_()_(),(),(_0_0?????? ?? ?yt e m pxt e m pjit e m pr a t eyxff i t n e s syt e mpjxt e mpi? (4) In , ),( jitemp is the gray level of the coordinates ),( ji in the template image,the width and height of which are xtemp_ and ytemp_ . ),( ?rateyxf gives the gray level in the original image,the coordinates of which are calculated by translation from),( yx ,and by changing the scale and the rotation angleθfrom the the images are 256 graylevel images,in ,division by 255 ensures that the resulting fitness is between 0 and maximum number of generations is limited to 300,and the threshold of the matching rate is set to That is to say,if within 300 generations the matching rate can reach ,then it is said that the template is found in the original image(the template matched the original image by the threshold).Otherwise,the result gives the best match until the training reaches the 300th generation. The results of GAbased face recognition are given in and Table 6a,c and d are searched to match the template ,while is matched to . Figure 6a and b reach the matching rate within 300 generations,while and d cannot reach the matching rate within 300 generations(the best match is given in Table 3).In the images in –c,we see that the result given matches the template coordinates),( yx ,the rate of scale,and the angle of rotationθhave been found correctly,but for ,the result is not very reason for this is that the template cannot represent the face of interest at all is to say, although the person to be recognized in different images is the same,the template cannot give the features for this person at all times(different appearance,etc.),and in all conditions.(The creation of the template is shown in .) A second reason is that the algorithm itself has some example,by using a GAbased recognition method,the settings of the search field(in this paper, ),( ?rateyx is selected),the determination of the geic operations,and the selection and optimization of the fitness function all have a strong effect on the level of recognition of the resultant image. of template For the purpose of paring the effects of the GAbased algorithm,the result of the general matching method7 is also
試題試卷相關(guān)推薦
鄂ICP備17016276號-1