Releases: pixpark/gpupixel
v1.2.x.8
iOS
工程根目录下执行如下命令
生成工程
bash [Arm64]
cmake -G Xcode -B build -S src -DCMAKE_TOOLCHAIN_FILE=../toolchain/ios.toolchain.cmake -DPLATFORM=OS64 -DCMAKE_BUILD_TYPE=Release
编译
bash [Release]
cmake --build build --config Release
bash [Debug]
cmake --build build --config Debug
Mac
工程根目录下执行如下命令
生成工程
bash [Apple Silicon]
cmake -G Xcode -B build -S src -DCMAKE_TOOLCHAIN_FILE=../toolchain/ios.toolchain.cmake -DPLATFORM=MAC_ARM64 -DCMAKE_BUILD_TYPE=Release
bash [Intel]
cmake -G Xcode -B build -S src -DCMAKE_TOOLCHAIN_FILE=../toolchain/ios.toolchain.cmake -DPLATFORM=MAC -DCMAKE_BUILD_TYPE=Debug
编译
bash [Release]
cmake --build build --config Release
[Debug]
cmake --build build --config Debug
Android
使用Android Studio打开目录 src/android/java, 开始自动下载 gradle 等依赖
工程结构
包含demo和 gpupixel module, 如下
![]()
编译
双击右侧 gradle -> gpupixel -> build -> assemble 开始编译
![]()
Windows
Windows编译需要安装 Cmake 和 MinGW64.
生成工程
bash [Release]
cmake -G "MinGW Makefiles" -B build -S src -DCMAKE_BUILD_TYPE=Release
bash [Debug]
cmake -G "MinGW Makefiles" -B build -S src -DCMAKE_BUILD_TYPE=Debug
编译
bash [Release]
cmake --build build --config Release
bash [Debug]
cmake --build build --config Debug
Linux (Test On Ubuntu)
环境配置
bash
install cmake
sudo apt-get install cmake pkg-config
install dependent lib
sudo apt-get install mesa-utils libglu1-mesa-dev freeglut3-dev mesa-common-dev libglfw3-dev
生成工程
bash [Release]
cmake -B build -S src -DCMAKE_BUILD_TYPE=Release
bash [Debug]
cmake -B build -S src -DCMAKE_BUILD_TYPE=Debug
编译
bash [Release]
cmake --build build --config Release
bash [Debug]
cmake --build build --config Debug
v1.3.1
v1.3.1
🛠️ Improvements
- MarsFace library and API updated with more accurate facial keypoints, reduced processing time and lower CPU usage;
- Windows and Linux platforms no longer provide built-in face detection, please integrate vnn if needed
- Android dynamic library supports 16k memory alignment
🐛 Bug Fixes
- Fixed iOS rendering black screen issue
🛠️ 改进 (Improvements)
- MarsFace 库和调用更新,人脸关键点更精准,耗时和 CPU 占用更小;
- Window 和 Linux 端不再提供内置人脸检测,有需要可以自行集成使用 vnn
- Android 动态库支持 16k 内存对齐
🐛 修复 (Bug Fixes)
- 修复 iOS 端渲染黑屏问题
v1.3.0-beta.7
- Fix Android crash caused by missing JNIEXPORT in JNI interface.
v1.3.0-beta.6
Version: v1.3.0-beta.6
What's New
- refine api
- refine demo
- bugfix
v1.3.0-beta5
GPUPixel Version: v1.3.0-beta5
What's New
- refine desktop demo
- gl thread run sync
v1.3.0-beta4
GPUPixel Version: v1.3.0-beta4
What's New
- refine api and demo
v1.3.0-beta3
GPUPixel Version: v1.3.0-beta3
What's New
- fix windows crash, can not find models file
- fix mac and ios framework not embeded resource file
- fix android demo crash
- change c++ standard to c++11
- update document and site
v1.3.0-beta
What's New
-
🔄 Replace the VNN face detection library with Mars-Face.
-
📉 Further reduce the library size.
-
🖥️ Use static linking for Mars-Face on Linux, macOS, and iOS.
-
📱 Use dynamic linking for Mars-Face on Windows and Android.
-
🐧 For Linux, the Mars-Face static library must be compiled using Debian 10.
v1.2.5
GPUPixel Version: v1.2.5
What's New
- Android support armeabi-v7a @cijiugechu f51ece9
- fix type errors in jni @cijiugechu f70fa25
- Refine android filter @xiaoshanlin000 7ee92c3
v1.2.4
GPUPixel Version: v1.2.4
What's New
- opt cmake and update demo