https://github.com/JuZiSYJ/centerface-ncnn

using ncnn to deploy centerface on Android
https://github.com/JuZiSYJ/centerface-ncnn

android centerface centernet facedetection ncnn

Last synced: 28 days ago
JSON representation

using ncnn to deploy centerface on Android

• Host: GitHub
• URL: https://github.com/JuZiSYJ/centerface-ncnn
• Owner: JuZiSYJ
• Created: 2020-06-02T13:34:58.000Z (almost 5 years ago)
• Default Branch: master
• Last Pushed: 2020-06-05T02:17:15.000Z (almost 5 years ago)
• Last Synced: 2024-08-01T05:12:27.594Z (9 months ago)
• Topics: android, centerface, centernet, facedetection, ncnn
• Language: C++
• Size: 1.13 MB
• Stars: 10
• Watchers: 2
• Forks: 2
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• awesome-ncnn - centerface-ncnn

README

        
# centerface-ncnn

using ncnn, on android,detection,surport **arm v7/v8/x86**, surport **vulkan** on gpu

## environment

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

AndroidStudio3.2(NDK19)  

Qualcomm845

## model

using pretrained **centerface** ,by[star-clouds](https://github.com/Star-Clouds/CenterFace)   

**pipeline**:  model on **PC -->   ONNX  ---> model on ncnn(.bin,,param) --> inference on Android**  

## cost_time(only once)

item | PC(i7 8700 500*300) |  Android 模拟 600*400 

-|-|-

load_model | 5ms | 35ms |

infer_time | 192ms | 190ms |

decode_time | 0ms | 1ms |

## organization

using **ncnn** and **opencv** on android  

the sim-overall android-organization is here

![organization](./test-img/buju.png)

## result

infer time on **845** in float32 is aroud ***350ms(700x500)***, depending on the object  

![845-cpu](./test-img/Screenshot_2020-06-02-21-13-29-537_com.example.sy.jpg)

![845-gpu](./test-img/Screenshot_2020-06-02-21-13-57-663_com.example.sy.jpg)

**ncnn** is good in optimazing on float32

## reference

(https://github.com/JuZiSYJ/MobilenetSSD_Android)

(https://github.com/nihui/ncnn-android-squeezenet)

(https://blog.csdn.net/qq_33431368/article/details/84990390)