[Dataset][VOC] Eyewear dataset 6000 in VOC format

Data set production unit: FIRC
Copyright unit: FIRC
Copyright statement: The data set is only for personal use, and organizations or individuals are not allowed to spread and conduct commercial activities on the Internet
Data set format: Pascal VOCFormat (excluding split txt files, only jpg images and corresponding xml)
Number of pictures (number of jpg files): 6121
Number of annotations (number of xml files): 6121

Authorization method: charge
Number of annotation categories: 2
Annotation category name: ["ng","g"]
Number of boxes marked in each category:
ng count = 3547
g count = 4745

Use the labeling tool: labelImg
Labeling rule: draw a rectangular frame for the category
Important note: g means wearing glasses, ng means not wearing glasses
Special statement: This dataset is not suitable for training models orThe accuracy of the weight file is guaranteed, and the dataset only provides accurate and reasonable annotations

Dataset reference address: Click me

Example dataset: