This tool is for object detection using rectangle box with methods like faster rcnn,forked from firstly faster rcnn, but improved for similarity search. the Second improvement is for segmentation like edge detection. The code is simply to create black image and draw white lines on it.

如果用图像标注工具labelme产生了xml文件,则将图像放到JPEGImages，xml文件放到Annotations下,tranval.txt放到ImageSets/Main下,运行show.py则在文件夹下BinaryImages下产生生成的二值图像