I want to train the masktrackrcnn, but it occur :KeyError: "YouTubeVISDataset: 'image_id'"

[eatbreakfast111]

Hello!
I want to train the masktrackrcnn by the official youtube_vis_dataset
but it occur :KeyError: "YouTubeVISDataset: 'image_id'".
Here is my datatree

Traceback (most recent call last):
  File "/home/music/miniconda3/envs/mmtrack/lib/python3.8/site-packages/mmcv/utils/registry.py", line 69, in build_from_cfg
    return obj_cls(**args)
  File "/home/music/Downloads/mmtracking-master/mmtrack/datasets/youtube_vis_dataset.py", line 44, in __init__
    super().__init__(*args, **kwargs)
  File "/home/music/Downloads/mmtracking-master/mmtrack/datasets/coco_video_dataset.py", line 46, in __init__
    super().__init__(*args, **kwargs)
  File "/home/music/miniconda3/envs/mmtrack/lib/python3.8/site-packages/mmdet/datasets/custom.py", line 97, in __init__
    self.data_infos = self.load_annotations(local_path)
  File "/home/music/Downloads/mmtracking-master/mmtrack/datasets/coco_video_dataset.py", line 61, in load_annotations
    data_infos = self.load_video_anns(ann_file)
  File "/home/music/Downloads/mmtracking-master/mmtrack/datasets/coco_video_dataset.py", line 73, in load_video_anns
    self.coco = CocoVID(ann_file)
  File "/home/music/Downloads/mmtracking-master/mmtrack/datasets/parsers/coco_video_parser.py", line 22, in __init__
    super(CocoVID, self).__init__(annotation_file=annotation_file)
  File "/home/music/miniconda3/envs/mmtrack/lib/python3.8/site-packages/mmdet/datasets/api_wrappers/coco_api.py", line 23, in __init__
    super().__init__(annotation_file=annotation_file)
  File "/home/music/miniconda3/envs/mmtrack/lib/python3.8/site-packages/pycocotools/coco.py", line 86, in __init__
    self.createIndex()
  File "/home/music/Downloads/mmtracking-master/mmtrack/datasets/parsers/coco_video_parser.py", line 57, in createIndex
    imgToAnns[ann['image_id']].append(ann)
KeyError: 'image_id'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "tools/train.py", line 213, in <module>
    main()
  File "tools/train.py", line 188, in main
    datasets = [build_dataset(cfg.data.train)]
  File "/home/music/miniconda3/envs/mmtrack/lib/python3.8/site-packages/mmdet/datasets/builder.py", line 82, in build_dataset
    dataset = build_from_cfg(cfg, DATASETS, default_args)
  File "/home/music/miniconda3/envs/mmtrack/lib/python3.8/site-packages/mmcv/utils/registry.py", line 72, in build_from_cfg
    raise type(e)(f'{obj_cls.__name__}: {e}')
KeyError: "YouTubeVISDataset: 'image_id'"

Thank you!

[dyhBUPT]

It seems that there is some error in your annotation file.

You can refer to:
https://github.com/open-mmlab/mmtracking/blob/master/docs/en/dataset.md#2-convert-annotations

[eatbreakfast111]

Thank you so much. It work.
I had write a script labelme2cocovid, but when i try to train the dataset, it occur this problem:

Here is my labelme2cocovid, could you please give me some advice?

#!/usr/bin/env python
# -*- coding:utf-8 -*-

import os
import cv2
import numpy as np
from tqdm import tqdm
import json

############################################################
# 存在图像和json的目录
srcpath = '/home/music/Downloads/mmtracking-master/data/youtube_vis_2019/valid/JPEGImages/'
# 要保存的结果json
dstjson = 'valid.json'
############################################################

# 读取图像，解决imread不能读取中文路径的问题
def cv_imread(filePath):
    cv_img = cv2.imdecode(np.fromfile(filePath,dtype=np.uint8),-1)
    return cv_img

if __name__ == '__main__':

    # 最终的json
    dst_ann = {
    "categories": [],
    "videos": [],
    "images": [],
    "annotations": []}
    categories_list = []
    
    # 遍历目录
    videos_list = os.listdir(srcpath)
    for videoname in videos_list:
        print('================='+videoname+'=================')
        videopath = srcpath + videoname + '/'
        fileslist = os.listdir(videopath)
        #############################################################
        # 写video相关信息
        curr_video_id = len(dst_ann['videos'])+1
        dst_ann['videos'].append({"id": curr_video_id,"name": videoname})
        #############################################################
        # 遍历一个文件夹下的图像及标注
        for filename in tqdm(fileslist):
            if '.json' in filename:
                jsonpath = videopath + filename
                imgpath = jsonpath[:-4] + 'jpg'

                img = cv_imread(imgpath)
                width,height,_ = img.shape

                #############################################################
                # 写images相关信息
                curr_img_id = len(dst_ann['images'])+1
                img_info =  {
                    "file_name": filename[:-4] + 'jpg', 
                    "height": height,"width": width,
                    "id": curr_img_id,
                    "video_id": curr_video_id,
                    "frame_id": int(filename[:-5]) }
                dst_ann['images'].append(img_info)
                #############################################################

                # 读json
                with open(jsonpath,encoding="utf-8") as f:
                    data = json.load(f)
                    shapes = data['shapes']
                    for s in shapes:
                        label = s['label']
                        points = s['points']
                        curr_category_id = -1
                        #############################################################
                        # 写类别相关信息
                        if not label in categories_list:
                            curr_category_id = len(categories_list)+1
                            categories_list.append(label)
                            dst_ann['categories'].append({"id": curr_category_id,"name": label})
                        else:
                            for c in dst_ann['categories']:
                                if c['name'] == label:
                                    curr_category_id = c['id']
                                    break
                        #############################################################
                        # 求点的外接矩形
                        xmin,xmax,ymin,ymax = width,0,height,0
                        for p in points:
                            x,y = p[0], p[1]
                            xmin = min(xmin, x)
                            xmax = max(xmax,x)
                            ymin = min(ymin, y)
                            ymax = max(ymax, y)
                        # check bbox
                        if 0:
                            cv2.rectangle(img, (int(xmin),int(ymin)), (int(xmax),int(ymax)), (255, 0, 255), 3)
                            cv2.imshow('',img)
                            cv2.waitKey(-1)
                        # 计算面积
                        area = (xmax-xmin) * (ymax-ymin)

                        #############################################################
                        # 写annotations相关信息 
                        curr_ann_id = len(dst_ann['annotations'])+1
                        ann = {"id": curr_ann_id,
                        "image_id": curr_img_id,
                        "video_id": curr_video_id,
                        "category_id": curr_category_id,
                        "instance_id": -1,
                        "bbox": [xmin,ymin,xmax-xmin,ymax-ymin],
                        "area": area,
                        "mask":points,
                        "occluded": False,
                        "truncated": False,
                        "iscrowd": False,
                        "ignore": False,
                        "is_vid_train_frame": True,
                        "visibility": 1.0}
                        dst_ann['annotations'].append(ann)
                        #############################################################
    #############################################################
    # 写结果文件              
    with open(dstjson,"w") as f:
        json.dump(dst_ann, f,
              indent=2,  # 空格缩进符，写入多行
              sort_keys=False,  # 键的排序
              ensure_ascii=True)  # 显示中文

Thank you!