A defect detection project with simple graphical interface based on pyqt.
Using ssd_mobilenet_v2_fpnlite_320x320_coco17_tpu-8 && efficientdet_d0_coco17_tpu-32 for model training on colab.
Based On:
https://github.com/nicknochnack/TFODCourse
Follow this tutorial:
https://youtu.be/yqkISICHH-U
Get dataset from NEU surface defect database(images with annotations):
http://faculty.neu.edu.cn/songkc/en/zhym/263264/list/index.htm
Get pre-trained object detection models(designed for tf2):
https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/tf2_detection_zoo.md
Or you can download those data directly here:
https://drive.google.com/open?id=1qrdZlaDi272eA79b0uCwwqPrm2Q_WI3k
https://pan.baidu.com/s/1nBbO-jWDm1_NHDQsc1dRkg (Code:pmqx)
For graduation thesis and project.