Knight/data_prepare_MSVD.py at main · junyangwang0410/Knight · GitHub

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
import random
import torch
from clip import clip
import numpy as np
import cv2
import os
from PIL import Image

def video2frames(videofile, savepath):
    vcap = cv2.VideoCapture()
    vcap.open(videofile)

    n = 1
    total_frames = int(vcap.get(cv2.CAP_PROP_FRAME_COUNT))
    frame_interval = int(total_frames / 5)

    for i in range(total_frames):
        _, frame = vcap.read()
        if i % frame_interval == 0:
            filename = videofile.split('.')[-1] + '_' + str(n) + '.jpg'

            cv2.imencode('.jpg', frame)[1].tofile(os.path.join(savepath, filename))
            n += 1
            if n == 6:
                break
    assert n == 6
    vcap.release()

device = "cuda:0"
clip_model, preprocess = clip.load("RN50x64", device = device)

with open("../Dataset/MSVD/caption.txt") as caption_data:
    captions_info = caption_data.readlines()

caption_dict = dict()

for caption_info in captions_info:
    video_name = caption_info.split(' ')[0]
    caption = caption_info[len(video_name)+1:].split('\n')[0]
    caption_dict.setdefault(video_name, []).append(caption)

with open("../Dataset/MSVD/train_list.txt") as train_data:
    train_names = train_data.readlines()

captions = []
for train_name in train_names:
    train_name = train_name.split('\n')[0]
    for caption in caption_dict[train_name]:
        captions.append(caption + '.')

random.shuffle(captions)

with open("../Dataset/MSVD/test_list.txt") as test_data:
    test_names = test_data.readlines()

with torch.no_grad():
    features = []
    index = 0
    batch_size = 256
    while index < len(captions):
        batch_captions = captions[index : index+batch_size]
        clip_captions = clip.tokenize(batch_captions).to(device)
        clip_features = clip_model.encode_text(clip_captions)
        features.append(clip_features)
        index += batch_size

    caption_features = torch.cat(features)

    torch.save(caption_features, "./feature/MSVD/caption_features.pkl")
    captions = np.array(captions)
    np.save("./feature/MSVD/captions.npy", captions)

    caption_features = caption_features / caption_features.norm(dim = -1, keepdim = True)

    nibers = []
    for i in range(caption_features.shape[0]):
        caption_feature = caption_features[i].unsqueeze(0)
        similarity = caption_feature @ caption_features.T
        similarity[0][i] = 0
        niber = []
        for j in range(5):
            _, max_id = torch.max(similarity, dim = 1)
            niber.append(max_id.item())
            similarity[0][max_id.item()] = 0

        nibers.append(niber)

    nibers = np.array(nibers)
    np.save("./feature/MSVD/nibers.npy", nibers)

    with open("./data/MSVD/test_list.txt") as video_name_data:
        video_name = video_name_data.readlines()

    for video_info in video_name:
        video_info = video_info.split('\n')[0]
        savepath = './feature/MSVD/frames/' + video_info + '/'
        os.mkdir(savepath)
        videofile = "./data/MSVD/video/" + video_info + '.avi'
        video2frames(videofile, savepath)

    path = "./feature/MSVD/frames/"
    with open("../Dataset/MSVD/test_list.txt") as video_name_data:
        video_names = video_name_data.readlines()

    video_features = []
    for video_name in video_names:
        video_name = video_name.split('\n')[0]
        video_path = path + video_name + '/'
        image_names = os.listdir(video_path)
        video_feature = []
        for image_name in image_names:
            ori_image = Image.open(video_path + image_name)
            image = preprocess(ori_image).unsqueeze(0).to(device)
            image_feature = clip_model.encode_image(image)
            video_feature.append(image_feature)
        video_feature = torch.cat(video_feature)
        video_features.append(video_feature.unsqueeze(0))
    video_features = torch.cat(video_features)

    torch.save(video_features, "./feature/MSVD/video_features.pkl")