File size: 1,500 Bytes
47af768
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
Taken from: https://competitions.codalab.org/competitions/20128#participate-get-data

The label file follows MSCOCO's style in json format. We adapt the entry name and label format for video. The definition of json file is:


        {
            "info" : info,
            "videos" : [video],
            "annotations" : [annotation],
            "categories" : [category],
        }
        video{
            "id" : int,
            "width" : int,
            "height" : int,
            "length" : int,
            "file_names" : [file_name],
        }
        annotation{
            "id" : int, 
            "video_id" : int, 
            "category_id" : int, 
            "segmentations" : [RLE or [polygon] or None], 
            "areas" : [float or None], 
            "bboxes" : [[x,y,width,height] or None], 
            "iscrowd" : 0 or 1,
        }
        category{
            "id" : int, 
            "name" : str, 
            "supercategory" : str,
        }
    
The submission file is also in json format. The file should contain a list of predictions:


        prediction{
            "video_id" : int, 
            "category_id" : int, 
            "segmentations" : [RLE or [polygon] or None], 
            "score" : float, 
        }
    
The submission file should be named as "results.json", and compressed without any subfolder. There is an example "valid_submission_sample.zip" in download links above. The example is generated by our proposed MaskTrack R-CNN algorithm.