English | 简体中文

AI Canteen

Faster and more convenient way of ordering based on Object detection

Android Live Demo inferenece of Yolov7 using ncnn

Object detection for 21 classes of dishes and ordering pages included

Install

Simply open this project with Android Studio, NCNN and OpenCV libraries are already included in this project

If you want to change the library version

• https://github.com/Tencent/ncnn/releases

1. Download ncnn-YYYYMMDD-android-vulkan.zip or build NCNN for android yourself
2. Extract ncnn-YYYYMMDD-android-vulkan.zip into app/src/main/jni and change the ncnn_DIR path to yours in app/src/main/jni/CMakeLists.txt

• https://github.com/nihui/opencv-mobile

1. Download opencv-mobile-XYZ-android.zip
2. Extract opencv-mobile-XYZ-android.zip into app/src/main/jni and change the OpenCV_DIR path to yours in app/src/main/jni/CMakeLists.txt

Custom Data

Train and convert

1. Training YOLOv7-tiny model
2. Convert the model to NCNN format into app/src/main/assets. I personally use PNNX

Mapping Array

Defined in app/src/main/java/com/tencent/ncnnyolov7/MenuActivity.java

f = new int[]{1, 2, 3, 15, 20, 21, 4, 22, 16, 5, 6, 7, 8, 9, 10, 17, 11, 12, 23, 18, 13};

This array solves the problem that the order page does not correspond to the classes number in the model

Details of Dishes

The dishes information is locally in the json string named eleme_json in app/src/main/res/values/strings.xml, and the picture is obtained by url

For more information please refer to Linkage-RecyclerView

Reference

Dataset: https://universe.roboflow.com/skoopin-data/pandadata

YOLOv7: https://github.com/WongKinYiu/yolov7

YOLOv7 for Android: https://github.com/xiang-wuu/ncnn-android-yolov7

Ordering page Architecture: https://github.com/KunMinX/Linkage-RecyclerView

License

GPL-3.0