这一版,对虹软的功能进行了一些封装,添加了人脸特征比对,比对结果保存到文件,和从文件提取特征进行比对,大体功能基本都已经实现,可以进行下一步的应用开发了
face_class.py
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
|
from ctypes import *#人脸框class MRECT(Structure): _fields_=[(u'left1',c_int32),(u'top1',c_int32),(u'right1',c_int32),(u'bottom1',c_int32)]#版本信息 版本号,构建日期,版权说明class ASF_VERSION(Structure): _fields_=[('Version',c_char_p),('BuildDate',c_char_p),('CopyRight',c_char_p)]#单人人脸信息 人脸狂,人脸角度class ASF_SingleFaceInfo(Structure): _fields_=[('faceRect',MRECT),('faceOrient',c_int32)]#多人人脸信息 人脸框数组,人脸角度数组,人脸数class ASF_MultiFaceInfo(Structure): # _fields_=[('faceRect',POINTER(MRECT)),('faceOrient',POINTER( c_int32)),('faceNum',c_int32)] _fields_=[(u'faceRect',POINTER(MRECT)),(u'faceOrient',POINTER(c_int32)),(u'faceNum', c_int32)] # _fields_=[(u'faceRect',MRECT*50),(u'faceOrient',c_int32*50),(u'faceNum',c_int32)]#人脸特征 人脸特征,人脸特征长度class ASF_FaceFeature(Structure): _fields_=[('feature',c_void_p),('featureSize',c_int32)]#自定义图片类class IM: def __init__(self): self.filepath=None self.date=None self.width=0 self.height=0 |
face_dll.py
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
|
from ctypes import *from face_class import *wuyongdll=CDLL('d:\python\Test\Face\lib\X64\libarcsoft_face.dll')dll=CDLL('d:\python\Test\Face\lib\X64\libarcsoft_face_engine.dll')dllc=cdll.msvcrtASF_DETECT_MODE_VIDEO = 0x00000000ASF_DETECT_MODE_IMAGE = 0xFFFFFFFFc_ubyte_p = POINTER(c_ubyte) #激活jihuo=dll.ASFActivationjihuo.restype = c_int32jihuo.argtypes = (c_char_p,c_char_p)#初始化chushihua=dll.ASFInitEnginechushihua.restype=c_int32chushihua.argtypes=(c_long,c_int32,c_int32,c_int32,c_int32,POINTER(c_void_p))#人脸识别shibie=dll.ASFDetectFacesshibie.restype=c_int32shibie.argtypes=(c_void_p,c_int32,c_int32,c_int32,POINTER(c_ubyte),POINTER(ASF_MultiFaceInfo))#特征提取tezheng=dll.ASFFaceFeatureExtracttezheng.restype=c_int32tezheng.argtypes=(c_void_p,c_int32,c_int32,c_int32,POINTER(c_ubyte),POINTER(ASF_SingleFaceInfo),POINTER(ASF_FaceFeature))#特征比对bidui=dll.ASFFaceFeatureComparebidui.restype=c_int32bidui.argtypes=(c_void_p,POINTER(ASF_FaceFeature),POINTER(ASF_FaceFeature),POINTER(c_float))malloc = dllc.mallocfree = dllc.freememcpy = dllc.memcpymalloc.restype = c_void_pmalloc.argtypes = (c_size_t, )free.restype = Nonefree.argtypes = (c_void_p, )memcpy.restype = c_void_pmemcpy.argtypes = (c_void_p, c_void_p, c_size_t) |
face_function.py
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
|
import face_dll,face_classfrom ctypes import *import cv2from io import BytesIO# from Main import *Handle=c_void_p()c_ubyte_p = POINTER(c_ubyte) # 激活函数def JH(appkey,sdkey): ret=face_dll.jihuo(appkey,sdkey) return ret# 初始化函数def CSH():# 1：视频或图片模式,2角度,3最小人脸尺寸推荐16,4最多人脸数最大50,5功能,6返回激活句柄 ret=face_dll.chushihua(0xFFFFFFFF,0x1,16,50,5,byref(Handle)) # Main.Handle=Handle return ret,Handle# cv2记载图片并处理def LoadImg(im): img=cv2.imread(im.filepath) sp=img.shape img=cv2.resize(img,(sp[1]//4*4,sp[0]//4*4)) sp=img.shape im.data=img im.width=sp[1] im.height=sp[0] return imdef RLSB(im): faces=face_class.ASF_MultiFaceInfo() img=im.data imgby=bytes(im.data) imgcuby=cast(imgby,c_ubyte_p) ret=face_dll.shibie(Handle,im.width,im.height,0x201,imgcuby,byref(faces)) return ret,faces# 显示人脸识别图片def showimg(im,faces): for i in range(0,faces.faceNum): ra=faces.faceRect[i] cv2.rectangle(im.data,(ra.left1,ra.top1),(ra.right1,ra.bottom1),(255,0,0,),2) cv2.imshow('faces',im.data) cv2.waitKey(0)#提取人脸特征def RLTZ(im,ft): detectedFaces=face_class.ASF_FaceFeature() img=im.data imgby=bytes(im.data) imgcuby=cast(imgby,c_ubyte_p) ret=face_dll.tezheng(Handle,im.width,im.height,0x201,imgcuby,ft,byref(detectedFaces)) if ret==0: retz=face_class.ASF_FaceFeature() retz.featureSize=detectedFaces.featureSize #必须操作内存来保留特征值,因为c++会在过程结束后自动释放内存 retz.feature=face_dll.malloc(detectedFaces.featureSize) face_dll.memcpy(retz.feature,detectedFaces.feature,detectedFaces.featureSize) # print('提取特征成功:',detectedFaces.featureSize,mem) return ret,retz else: return ret#特征值比对,返回比对结果def BD(tz1,tz2): jg=c_float() ret=face_dll.bidui(Handle,tz1,tz2,byref(jg)) return ret,jg.value#单人特征写入文件def writeFTFile(feature,filepath): f = BytesIO(string_at(feature.feature,feature.featureSize)) a=open(filepath,'wb') a.write(f.getvalue()) a.close()#从多人中提取单人数据def getsingleface(singleface,index): ft=face_class.ASF_SingleFaceInfo() ra=singleface.faceRect[index] ft.faceRect.left1=ra.left1 ft.faceRect.right1=ra.right1 ft.faceRect.top1=ra.top1 ft.faceRect.bottom1=ra.bottom1 ft.faceOrient=singleface.faceOrient[index] return ft#从文件获取特征值def ftfromfile(filepath): fas=face_class.ASF_FaceFeature() f=open('d:/1.dat','rb') b=f.read() f.close() fas.featureSize=b.__len__() fas.feature=face_dll.malloc(fas.featureSize) face_dll.memcpy(fas.feature,b,fas.featureSize) return fas |
Main1.py
|
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
|
import face_dll,face_classfrom ctypes import *import cv2import face_function as funAppkey=b''SDKey=b''# 激活ret=fun.JH(Appkey,SDKey)if ret==0 or ret==90114: print('激活成功:',ret)else: print('激活失败:',ret) pass# 初始化ret=fun.CSH()if ret[0]==0: print('初始化成功:',ret,'句柄',fun.Handle)else: print('初始化失败:',ret)# 加载图片im=face_class.IM()im.filepath='e:/2.jpg'im=fun.LoadImg(im)print(im.filepath,im.width,im.height)# cv2.imshow('im',im.data)# cv2.waitKey(0)print('加载图片完成:',im)ret=fun.RLSB(im)if ret[0]==-1: print('人脸识别失败:',ret) passelse: print('人脸识别成功:',ret)# 显示人脸照片# showimg(im,ret)#提取单人1特征ft=fun.getsingleface(ret[1],0)tz1=fun.RLTZ(im,ft)[1]#提取单人2特征ft=fun.getsingleface(ret[1],1)tz2=fun.RLTZ(im,ft)[1]#特征保存到文件# fun.writeFTFile(tz1,'d:/1.dat')# fun.writeFTFile(tz2,'d:/2.dat')#文件获取特征tz=fun.ftfromfile('d:/1.dat')jg=fun.BD(tz1,tz)print(jg[1])#结果比对# jg=fun.BD(tz1,tz2)# print(jg[1]) |
以上就是本文的全部内容，希望对大家的学习有所帮助，也希望大家多多支持服务器之家。
原文链接：https://www.cnblogs.com/wxt51/p/10125460.html
