【发布时间】：2021-02-25 16:39:06
【问题描述】：
我一直在使用 VGG Image Annotator 使用 GitLab 的本地版本 (via-3.0.9) 来注释图像。我已将图像导出为 CSV 文件，但在重新启动计算机并返回 VIA 后，我无法导入 CSV 文件，因为它仅支持用于注释导入的 JSON 格式。我也尝试过允许 CSV 导入的在线 VIA 应用程序，但我第一次尝试时，出现了一些关于缺少标题的错误，随后的导入尝试什么也没做。
导出的样子：
# Exported using VGG Image Annotator (http://www.robots.ox.ac.uk/~vgg/software/via)
# Notes:
# - spatial_coordinates of [2,10,20,50,80] denotes a rectangle (shape_id=2) of size 50x80 placed at (10,20)
# - temporal coordinate of [1.349,2.741] denotes a temporal segment from time 1.346 sec. to 2.741 sec.
# - temporal coordinate of [4.633] denotes a still video frame at 4.633 sec.
# - metadata of {""1"":""3""} indicates attribute with id "1" set to an attribute option with id "3"
# SHAPE_ID = {"POINT":1,"RECTANGLE":2,"CIRCLE":3,"ELLIPSE":4,"LINE":5,"POLYLINE":6,"POLYGON":7,"EXTREME_RECTANGLE":8,"EXTREME_CIRCLE":9}
# FLAG_ID = {"RESERVED_FOR_FUTURE":1}
# ATTRIBUTE = {}
# CSV_HEADER = metadata_id,file_list,flags,temporal_coordinates,spatial_coordinates,metadata
"1_auhS9dP3","[""train_in_9344_256.png""]",0,"[]","[6,13.474,1.078,12.665,68.985,15.629,98.897,14.552,194.829,21.019,203.992,12.935,213.962,12.935,221.777,19.402,247.107]"
"1_2CrCZ69X","[""train_in_9344_256.png""]",0,"[]","[6,127.731,1.617,128.269,246.029,128.539,253.575]","{}"
"1_oGrbStD2","[""train_in_9344_256.png""]",0,"[]","[6,194.56,253.844,188.901,210.728,179.469,147.402]","{}"
是否可以将其导入回 VIA 中？
【问题讨论】：
标签： computer-vision