Project Highlights:
- Target detection of mask using SSD algorithm
- The data set used can be made by yourself
An SSD-based mask wearing detection project
SSD-based mask wearing detection project
- Project Introduction
- Review of SSD Algorithm Principles
- data set collection
- Ultra-simple SSD algorithm implementation
- Ordinary SSD algorithm implementation
- Project source code interpretation
Two SSD algorithm
Single Shot MultiBox Detector
SSD Network is a paper published by author Wei Liu on ECCV 2016 . For a network with an input size of 300x300, use Nvidia Titan X to achieve 74.3%mAP and 59FPS on the VOC 2007 test set. For a 512x512 network , it reached 76.9%mAP , surpassing the strongest Faster RCNN (73.2%mAP) at that time .
2.1 Problems with Faster RCNN
- Poor detection of small objects
- The model is large and the detection speed is slow
2.2 Predicting targets
2.3 V GG- 16 model
2.4 Single Shot MultiBox Detector
2.5 Default Box scale and aspect settings _
2.6 Calculation of loss
2.6.1 Class loss
2.6.2 Localization loss