This is an Android Application aimed to make your community safer by implementing the SOTA deep neural network onto x86_64 or arm architecture. The application has two functions up to now, i.e. Object Detection via YOLOv5, and Facial Recognition via ScrFD.
Once you launch the application, you'll see the main page of object detection. Firstly, you need to put a picture onto the canvas to let the algorithm to detect it. You can either press album button to select a photo from the system album, or press the camera button to take a new picture with your device's camera. As soon as the photo shows up on the page, you can then press the detect button to detect the object on your selected picture.
As for the face recognition module, just enter the second page and you'll see the video captured by your front or back camera, the algorithm detects faces in the picture automatically.
This project was done by myself during my Applied Machine Learning SDE internship at Lenovo, the application is still looking forward to updates currently.
- Object Detection: YOLOv5
- Facial Recognition: ScrFD
Tencent ncnn