hog特征+svm分类器识别猫

hog特征+svm分类器识别猫

import cv2
import numpy as np
import matplotlib.pyplot as plt
import os

PosNum = 300
NegNum = 1000
winSize = (640,320)
blockSize = (80,80)
blockStride = (40,40)
cellSize = (40,40)
nBin = 9

#hog creat
hog = cv2.HOGDescriptor(winSize,blockSize,blockStride,cellSize,nBin)
#svm creat
svm = cv2.ml.SVM_create()
#couputer hog
featureNum = int(((320-80)/40+1)*((640-80)/40+1)*4*9)
featureArray = np.zeros(((PosNum+NegNum),featureNum),np.float32)
labelArray = np.zeros(((PosNum+NegNum),1),np.int32)

print(featureArray)
print(labelArray)

for i in range(0,PosNum):
    os.chdir(r"C:\Users\lenovo\Desktop\python\jiqixuexi\mao_shibie\pos_yangben")
    fileName = 'image' + str(i+1) + '.jpg'
    img = cv2.imread(fileName)
    hist = hog.compute(img,(40,40))  #3780
    for j in range(0,featureNum):
        featureArray[i,j] = hist[j]
    labelArray[i,0] = 1
    
for i in range(0,NegNum):
    os.chdir(r"C:\Users\lenovo\Desktop\python\jiqixuexi\mao_shibie\neg_yangben")
    fileName = 'image' +str(i+1) + '.jpg'
    hist = hog.compute(img,(40,40))
    for j in range(0,featureNum):
        featureArray[i+PosNum,j] = hist[j]
    labelArray[i+PosNum,0] = -1

#shuxingshezhi
svm.setType(cv2.ml.SVM_C_SVC)
svm.setKernel(cv2.ml.SVM_LINEAR)
svm.setC(0.01)

#TRAIN

print('开始训练')
ret = svm.train(featureArray,cv2.ml.ROW_SAMPLE,labelArray)
print('训练完成')

os.chdir(r"C:\Users\lenovo\Desktop\python\jiqixuexi\mao_shibie")
svm.save("SVM_HOG.xml")

alpha = np.zeros((1),np.float32)
rho = svm.getDecisionFunction(0,alpha)
print(rho)
print(alpha)
alphaArray = np.zeros((1,1),np.float32)
supportVArray = np.zeros((1,featureNum),np.float32)
resultArray = np.zeros((1,featureNum),np.float32)
alphaArray[0,0] = alpha
resultArray = -1*alphaArray*supportVArray

myDetect = np.zeros((3781),np.float32)
for i in range(0,3780):
    myDetect[i] = resultArray[0,i]
myDetect[3780] = rho[0]
myHog = cv2.HOGDescriptor()
myHog.setSVMDetector(myDetect)

imageSrc = cv2.imread(r"C:\Users\lenovo\Desktop\python\jiqixuexi\mao_shibie\pos_yangben\image9.jpg",1)
objs = myHog.detectMultiScale(imageSrc,0,(8,8),(32,32),1.05,2)

x = int(objs[0][0][0])
y = int(objs[0][0][1])
w = int(objs[0][0][2])
h = int(objs[0][0][3])
        
cv2.rectangle(imageSrc,(x,y),(x+w,y+h),(255,0,0),2)
cv2.imshow("dstt",imageSrc)
cv2.waitKey(0)

效果很不理想，后期继续改进。

注意：

可以保存训练好的特征和分类器，保存为xml，后边直接调用就可以了。

svm.save("SVM_HOG.xml")

参考链接：

OPENCV HOG特征+SVM分类器行人识别（从训练到识别）
https://blog.csdn.net/u011263315/article/details/41447545?utm_medium=distribute.pc_relevant_download.none-task-blog-baidujs-1.nonecase&depth_1-utm_source=distribute.pc_relevant_download.none-task-blog-baidujs-1.nonecase