Найти некоторый объект на изображении

import cv2 as cv
 
def main(filename='image.png'):
    image = cv.imread(filename, cv.IMREAD_GRAYSCALE)
    hog = cv.HOGDescriptor()
    features = hog.compute(image)
    print(features)
    pass
 
if __name__ == '__main__':
    main()
else:
    pass

import numpy as np
import cv2
cascade = cv2.CascadeClassifier("haarcascade_frontalface_alt.xml")
cap = cv2.VideoCapture(0)
cv2.namedWindow("camera")
#ret = cap.set(3,320)
ret = cap.set(4,240)
#ret = cap.set(5,15)
while(True):
    ret, frame = cap.read()
    cv2.imshow("camera", frame)
    gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
    faces = cascade.detectMultiScale(gray, 1.3, 4, cv2.cv.CV_HAAR_SCALE_IMAGE, (20,20))
    if len(faces) > 0:
        for (x,y,w,h) in faces:
            cv2.rectangle(frame, (x,y), (x+w,y+h), (255, 0, 0))
    cv2.imshow("camera", frame)
    if cv2.waitKey(20) == 27:
        break
cap.release()
cv2.destroyAllWindows()

#! /usr/bin/env plabelsthon2
 
import cv2 as cv
import sys
from contextlib import contextmanager
from multiprocessing.pool import ThreadPool
from StringIO import StringIO
from svmutil import *
 
GLOBAL_MODEL_FILENAME = 'model.dat'
GLOBAL_TRAINING_DATA_FILENAME = 'train.dat'
GLOBAL_TEST_DATA_FILENAME = 'test.dat'
GLOBAL_PATCH_WIDTH = 80
GLOBAL_PATCH_HEIGHT = 200
GLOBAL_WINDOW_WIDTH = 64
GLOBAL_WINDOW_HEIGHT = 128
GLOBAL_MAX_THREADS_COUNT = 8
 
 
@contextmanager
def nooutput():
    """Suppresses stdout."""
    saved_stdout = sys.stdout
    sys.stdout = StringIO()
    yield
    sys.stdout = saved_stdout
 
 
def build_hog(args):
    """Builds HOG for image given."""
    filename, label = args
    image = cv.imread(filename, cv.IMREAD_GRAYSCALE)
    image = cv.resize(image[:GLOBAL_PATCH_HEIGHT, :GLOBAL_PATCH_WIDTH],
                      (GLOBAL_WINDOW_WIDTH, GLOBAL_WINDOW_HEIGHT))
    hog = cv.HOGDescriptor()
    return float(label), [e for s in hog.compute(image) for e in s]
 
 
def train(filename=GLOBAL_TRAINING_DATA_FILENAME):
    """Performs classifier training."""
    with open(filename, 'r') as f:
        data = [tuple(e.split()) for e in f.read().split('\n')][:-1]
    print('Processing images...')
    pool = ThreadPool(GLOBAL_MAX_THREADS_COUNT)
    result = pool.map_async(build_hog, data)
    data = result.get()
    y = [y for y, _ in data]
    x = [x for _, x in data]
    assert(len(y) == len(x))
    print('Training SVM...')
    model = svm_train(y, x, '-t 0 -c 2 -e 0.001')
    return model
 
 
def test(model, filename=GLOBAL_TEST_DATA_FILENAME):
    """Performs classifier test using model and test data given."""
    y, x = svm_read_problem(filename)
    print('Testing model...')
    svm_predict(y, x, model)
    pass
 
 
def load_model(filename=GLOBAL_MODEL_FILENAME):
    """Loads model from file."""
    with nooutput():
        model = svm_load_model(filename)
    assert model is not None
    return model
 
 
def save_model(model, filename=GLOBAL_MODEL_FILENAME):
    """Saves model to file."""
    svm_save_model(filename, model)
    pass
 
 
def main():
    """Entry point."""
    try:
        model = load_model()
    except:
        print("Model not found, training required.")
        model = train()
        save_model(model)
    else:
        print("Model loaded.")
    test(model)
    pass
 
if __name__ == '__main__':
    main()
else:
    pass