voc2coco
voc2coco copied to clipboard
Tony607
How to create custom COCO data set for object detection
I encountered the error below so I fixed it. `ValueError: invalid literal for int() with base 10: '45.70000076293945'`
Number of xml files: 3000 Traceback (most recent call last): File "E:/voc2coco-master/voc2coco.py", line 157, in convert(xml_files, json_file) File "E:/voc2coco-master/voc2coco.py", line 83, in convert filename = get_and_check(root, "filename", 1).text File "E:/voc2coco-master/voc2coco.py",...
fixes the following: > - Requirement for the file names to be integers. > - Requirement for the co-ods to be integers.
Float in the coordinate will cause error: ``` Traceback (most recent call last): File "voc2coco.py", line 154, in convert(xml_files, args.json_file) File "voc2coco.py", line 109, in convert ymin = int(get_and_check(bndbox, "ymin",...
function 'get_filename_as_int' changed to 'get_filename' to reflect change of integer requirement
Category index now starts from 1 instead of 0
COCO expect that class ID starts from 1 as below, https://github.com/Tony607/voc2coco/blob/207f265dfc4afaae0dda229b8c27a91ab7e0494a/voc2coco.py#L14 However, It came to the notice that class ID starts from 0 in [output.json](https://github.com/Tony607/voc2coco/blob/master/data/coco/output.json). To fix this issue, change...
Hello, why is the JSON file name I generated called BMP?My picture turned out to be JPG.for example "file_name": "030003592.bmp",
The following changes made: * compute segments from bounding box. * make img_id and anno_id independent with voc annotation, the original way may cause some bugs if voc annotation file...
