This project is the first working version of Waifu2x ncnn Android app, which was initially developed from ncnn-android-deeplabv3plus as a template and reference the build method from ncnn-android-yolov5. And it has been updated with the latest JNI binding and build methods which is currently used by Waifu2x ncnn Android.
I hope that this project can be used as a reference by other developers to build more Android AI apps on Waifu2x ncnn Vulkan, or even more ncnn projects.
Just like other ncnn android demo projects, the only preparation is just about putting the approperiate ncnn lib into app/src/main/jni . However, in Waifu2x ncnn Android, there are mainly two options to choose from, and these result in different ncnn lib requirements. These two options are CMake flags: "USE_PREBUILT_NCNN" and "USE_SHARED_NCNN".
| \ | USE_PREBUILT_NCNN=ON (Default) | USE_PREBUILT_NCNN=OFF |
|---|---|---|
| USE_SHARED_NCNN=ON | ncnn-android-vulkan-shared | ncnn (git repo) |
| USE_SHARED_NCNN=OFF (Default) | ncnn-android-vulkan | ncnn (git repo) |
To choose the ncnn build type, modify the corresponding section in the app: build.gradle
- Sometimes, build may not be successful. (.etc, waifu2x.h cannot found VkVulkan syntex).
I feel very appreciated to the following people and projects.
- nihui: Helped me to solve the critical issue in the JNI of Waifu2x ncnn Vulkan at the ever beginning of this project.
- waifu2x-ncnn-vulkan: The original ncnn project of Waifu2x.
- ncnn-android-yolov5: Provides the right way to use the current prebuilt ncnn library.
- ncnn-android-deeplabv3plus: work as a template to build ncnn android app.