Android app for comparing traditional and learned techniques of feature detection and description. It provides a GUI for selecting a keypoint detection algorithm and displays an image with keypoints and their detection time in soft real-time.

Supported algorithms:

App uses its original library consisting of parts: native (traditional algorithms from OpenCV) and JVM (learned algorithms based on PyTorch for Android).

Follow these steps to buils the app:

1. Configure traditional algorithms: run BuildOpenCV script to automatically download and install OpenCV for the required Android ABI. Read more about this script and its arguments in the native README.
2. Configure learned algorithms:

# In any of the the project directories pip install pytorch # Installs PyTorch git submodule update --init # Initializes SuperPoint Git submodule # In `utils` subdirectory python save_for_mobile.py # Converts SuperPoint to mobile and saves it to library's assets

3. Build the project with Gradle for the required ABI (this must be the same ABI you built OpenCV for).

The app is very easy to use: give it the camera permission, and you are free to analyze everything around you with it!

• To select a keypoint detection algorithm, open the menu with a swipe to the right.
• By default, the None mode is set, in which only the image from the camera is displayed on the screen.
• Select the desired algorithm, close the menu, and point the camera at the scene of interest and observe.
• The obtained keypoints and their detection time will be displayed on screen.