svm_live_video_train_ex01.py

import numpy as np
import cv2
import cv
import sys
import os



video_capture = cv2.VideoCapture(0)
SZ=20
bin_n = 16 # Number of bins


samples= np.empty((0,2500))
responses=[]



class StatModel(object):
    '''parent class - starting point to add abstraction'''    
    def load(self, fn):
        self.model.load(fn)
    def save(self, fn):
        self.model.save(fn)



class SVM(StatModel):
    '''wrapper for OpenCV SimpleVectorMachine algorithm'''
    def __init__(self):
        self.model = cv2.SVM()

    def train(self, samples, responses):
        #setting algorithm parameters
        params = dict( kernel_type = cv2.SVM_LINEAR, 
                       svm_type = cv2.SVM_C_SVC,
                       C = 1 )
        self.model.train(samples, responses, params = params)

    def predict(self, samples):
        return np.float32( [self.model.predict(s) for s in samples])
    def predict2(self, samples):
        return np.float32( self.model.predict(samples))

def hog(img):
    gx = cv2.Sobel(img, cv2.CV_32F, 1, 0)
    gy = cv2.Sobel(img, cv2.CV_32F, 0, 1)
    mag, ang = cv2.cartToPolar(gx, gy)
    # quantizing binvalues in (0...16)
    bins = np.int32(bin_n*ang/(2*np.pi))

    # Divide to 4 sub-squares
    bin_cells = bins[:10,:10], bins[10:,:10], bins[:10,10:], bins[10:,10:]
    mag_cells = mag[:10,:10], mag[10:,:10], mag[:10,10:], mag[10:,10:]
    hists = [np.bincount(b.ravel(), m.ravel(), bin_n) for b, m in zip(bin_cells, mag_cells)]
    hist = np.hstack(hists)
    return hist


# keyboard mappings for 0-9; user may type in this range when prompted
keys = [i for i in range(48, 58)]


clf = SVM()
#clf.train(np.asarray(samples,dtype=np.float32), np.asarray(responses,dtype=np.float32))

if __name__ == '__main__':
   
    #global samples
    
    res=[]
    dfi=0
    
    
    while True:
        #=============================================================
        # Capture frame-by-frame
        #--------------------------------------------------------------
        
        ret, frame = video_capture.read()


        #===================================
        #do all processing Here
        #-----------------------------------

        #pre processing
        #prepare image for motion Detection
        #--------------------------------------

        gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
        gray = cv2.GaussianBlur(gray, (21, 21), 0)

        cv2.rectangle(frame,(300,300),(100,100),(0,255,0),0)
        crop_img = frame[100:300, 100:300]

        sil_gray = cv2.cvtColor(crop_img, cv2.COLOR_BGR2GRAY)
        kernel = (25, 25)
        sil_blur = cv2.GaussianBlur(sil_gray, kernel, 0)

        _, sil_bin = cv2.threshold(sil_blur, 127, 255,
                               cv2.THRESH_BINARY_INV+cv2.THRESH_OTSU)


        roi_small=cv2.resize(sil_bin,(50,50))
        roi_q = roi_small.reshape((1,2500))
        roi_q = np.float32(roi_q)


        if dfi>5:
            y_val = clf.predict(roi_q)
            print y_val
        #print y_val


        hog_hist=hog(sil_bin)
        #print hog_hist

        


        #==============================================
        #display area
        #----------------------------------------------

        cv2.imshow('Video', frame)
        cv2.imshow('vid',sil_bin)
        #cv2.imshow('hist',hog_hist)
        k=cv2.waitKey(1)
        
        if k & 0xFF == ord('q'):            
            break
        elif k in keys:
            
            if hog_hist.any():
                sp = hog_hist.reshape(1,64)
                samples = np.append(samples,roi_q,0)
                responses.append(int(chr(k)))
                print samples
                dfi=dfi+1
                if dfi> 2:
                    clf.train(np.asarray(samples,dtype=np.float32), np.asarray(responses,dtype=np.float32))
                #res = res.append(int(chr(k)))
                
               # responses=responses.append(int(chr(k)))
               


    # When everything is done, release the capture
    video_capture.release()
    cv2.destroyAllWindows()
#========================================================================================