Super-Resolution Software

Features

The software includes a motion estimation, or geometry registration. The registration process is automatically done. The software provides users option to provide the motion information estimated independently.
For the geometry registration, users can chose the motion models from translation, affine warp, or projective warp.
Users can check the super-resolution result through GUI. The motion estimation results are also appeared visually and numerically.
Users are able to specify the region of interest to be processed. Users are also able to specify the reference frame.
The software supports BMP file format for input and output.
The software generates a color high-resolution image from raw data of the camera with color filter array.
Users can configurate parameters for super-resolution process via GUI. Box PSF (Point Spread Function) and Gaussian PSF are available. Size of both PSFs can be configurable. Other typical parameters are resolution enhancement ratio, constraint parameter, constraint kernel, etc.
The super-resolution software has a capability to call image-capturing software. Observed images by the image-capturing software are directly proccessed for resolution enhancement.


Observed image (60x60)	Super-resolution result (240x240)
Using 30 frames, process is done in 4.2 second. This time includes capturing, registration, and super-resolution process.


Observed image (60x60)	Super-resolution result (180x180)
Using 10 frames, process is done in 1.9 second including capturing time. This time includes capturing, registration, and super-resolution process.