https://github.com/mlbo/insightface-npu

https://github.com/mlbo/insightface-npu

Last synced: 12 days ago
JSON representation

• Host: GitHub
• URL: https://github.com/mlbo/insightface-npu
• Owner: mlbo
• Created: 2021-08-13T02:38:28.000Z (over 3 years ago)
• Default Branch: master
• Last Pushed: 2021-08-27T05:45:55.000Z (about 3 years ago)
• Last Synced: 2023-03-06T17:43:01.676Z (over 1 year ago)
• Language: C++
• Size: 44.3 MB
• Stars: 19
• Watchers: 1
• Forks: 3
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Insightface Android NPU

## [app/src/main](https://github.com/mlbo/Insightface-NPU/tree/master/app/src/main)目录结构

```

├─assets                                存放模型

├─java                                  Java代码

├─jni                                   C++代码

│  ├─algorithm                          SCRFD代码

│  ├─include                            

│  │  └─image                           图像处理代码

│  ├─opencv-mobile-3.4.15-android       OpenCV mobile     

│  │  

│  └─utility                            SCRFD涉及处理代码

├─jniLibs

│  ├─arm64-v8a                          64位Tengine Android动态库

│  └─armeabi-v7a                        32位支持TIM-VX的Tengine Android动态库

```

## 准备

0. Khadas VIM3开发板一套

1. 用 Android Studio 建立一个空白Android项目

2. 把按照[交叉编译 Arm32/64 Android 版本](https://github.com/OAID/Tengine/blob/tengine-lite/doc/compile.md#3-%E4%BA%A4%E5%8F%89%E7%BC%96%E8%AF%91-arm3264-android-%E7%89%88%E6%9C%AC)和[编译 Android 32bit 平台](https://github.com/OAID/Tengine/blob/tengine-lite/doc/npu_tim-vx_user_manual_zh.md#29-%E7%BC%96%E8%AF%91-android-32bit-%E5%B9%B3%E5%8F%B0)

教程编译好的so分别复制到Android项目中jniLibs目录中arm64-v8a和armeabi-v7a文件夹内

3. 把[SCRFD模型](#)模型复制到Android项目中assets目录

4. 把[Tengine/demos/](https://github.com/OAID/Tengine/tree/tengine-lite/demos0)中的SCRFD涉及的代码复制到Android项目中的jni目录

## 编译环境搭建

参考[Insightface-NPU/app/build.gradle](https://github.com/mlbo/Insightface-NPU/blob/master/app/build.gradle)

### Android SDK

```

  minSdkVersion 24

  targetSdkVersion 30

```

### Android NDK

```

  ndk {

      moduleName "tengine-lite"

      abiFilters "armeabi-v7a",'arm64-v8a'

      ndkVersion "21.4.7075529"

  }

```

### Android OpenCV

下载[opencv-mobile-3.4.15-android.zip](https://github.com/nihui/opencv-mobile/releases/download/v13/opencv-mobile-3.4.15-android.zip)，然后解压放入jni目录

## TODO

- [ ] 针对Android手机性能优化

## 参考

1. [Tengine](https://github.com/OAID/Tengine)

2. [ncnn](https://github.com/Tencent/ncnn)

3. [opencv-mobile](https://github.com/nihui/opencv-mobile)

4. [ncnn-android-nanodet](https://github.com/nihui/ncnn-android-nanodet)