File size: 4,335 Bytes
9f3fa29
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
# -*- coding: utf-8 -*-
"""
Created on Thu Jun 24 11:36:01 2021

@author: Xinya
"""

import os
import glob
import time
import numpy as np
import csv
import cv2
import dlib

from skimage import transform as tf

detector = dlib.get_frontal_face_detector()
predictor = dlib.shape_predictor('./shape_predictor_68_face_landmarks.dat')


import imageio



def save(path, frames, format):
    if format == '.mp4':
        imageio.mimsave(path, frames)
    elif format == '.png':
        if not os.path.exists(path):


            os.makedirs(path)
        for j, frame in enumerate(frames):
            cv2.imwrite(path+'/'+str(j)+'.png',frame)
    #        imageio.imsave(os.path.join(path, str(j) + '.png'), frames[j])
    else:
        print ("Unknown format %s" % format)
        exit()

def crop_image(image_path, out_path):
    template = np.load('./M003_template.npy')
    image = cv2.imread(image_path)
    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
    rects = detector(gray, 1)  #detect human face
    if len(rects) != 1:
        return 0
    for (j, rect) in enumerate(rects):
        shape = predictor(gray, rect) #detect 68 points
        shape = shape_to_np(shape)

    pts2 = np.float32(template[:47,:])
    # pts2 = np.float32(template[17:35,:])
    # pts1 = np.vstack((landmark[27:36,:], landmark[39,:],landmark[42,:],landmark[45,:]))
    pts1 = np.float32(shape[:47,:]) #eye and nose
    # pts1 = np.float32(landmark[17:35,:])
    tform = tf.SimilarityTransform()
    tform.estimate( pts2, pts1) #Set the transformation matrix with the explicit parameters.
    
    dst = tf.warp(image, tform, output_shape=(256, 256))

    dst = np.array(dst * 255, dtype=np.uint8)
    
    
    cv2.imwrite(out_path,dst)

def shape_to_np(shape, dtype="int"):
    # initialize the list of (x, y)-coordinates
    coords = np.zeros((shape.num_parts, 2), dtype=dtype)

    # loop over all facial landmarks and convert them
    # to a 2-tuple of (x, y)-coordinates
    for i in range(0, shape.num_parts):
        coords[i] = (shape.part(i).x, shape.part(i).y)

    # return the list of (x, y)-coordinates
    return coords


def crop_image_tem(video_path, out_path):
    image_all = []
    videoCapture = cv2.VideoCapture(video_path)
    success, frame = videoCapture.read()
    n = 0
    while success :
        image_all.append(frame)
        n = n + 1
        success, frame = videoCapture.read()
        
    if len(image_all)!=0 :
        template = np.load('./M003_template.npy')
        image=image_all[0]
        gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
        rects = detector(gray, 1)  #detect human face
        if len(rects) != 1:
            return 0
        for (j, rect) in enumerate(rects):
            shape = predictor(gray, rect) #detect 68 points
            shape = shape_to_np(shape)

        pts2 = np.float32(template[:47,:])
        # pts2 = np.float32(template[17:35,:])
        # pts1 = np.vstack((landmark[27:36,:], landmark[39,:],landmark[42,:],landmark[45,:]))
        pts1 = np.float32(shape[:47,:]) #eye and nose
        # pts1 = np.float32(landmark[17:35,:])
        tform = tf.SimilarityTransform()
        tform.estimate( pts2, pts1) #Set the transformation matrix with the explicit parameters.
        out = []
        for i in range(len(image_all)):
            image = image_all[i]
            dst = tf.warp(image, tform, output_shape=(256, 256))

            dst = np.array(dst * 255, dtype=np.uint8)
            out.append(dst)
        if not os.path.exists(out_path):
            os.makedirs(out_path)
        save(out_path,out,'.png')

def proc_audio(src_mouth_path, dst_audio_path):
    audio_command = 'ffmpeg -i \"{}\" -loglevel error -y -f wav -acodec pcm_s16le ' \
                    '-ar 16000 \"{}\"'.format(src_mouth_path, dst_audio_path)
    os.system(audio_command)



if __name__ == "__main__":
    #video alignment
    video_path = './test/crop/M030_sad_3_001.mp4'
    out_path = './test/crop/M030_sad_3_001'
    crop_image_tem(video_path, out_path)
    
    #image alignment
    image_path = './test/raw_image/brade2.jpg'
    out_path = './test/image/brade2.jpg'
    crop_image(image_path, out_path)
    
    #change_audio_sample_rate
    src_mouth_path = './test/audio/00015.mp3'
    dst_audio_path = './test/audio/00015.mov'
    proc_audio(src_mouth_path, dst_audio_path)