MotionDSP's patented super-resolution algorithms have the unique ability reconstruct video, increasing resolution and removing noise. The super-resolution reconstruction process involves analyzing a large number of video frames, determining what is the best information, and using that information to reconstruct each frame of video.
The Super-resolution Reconstruction Process
A video is a series of snapshots of a scene. Due to camera or subject motion, each video frame provides a slightly different view, collectively providing a rich amount of information about the scene. MotionDSP's algorithms first accurately determine the motion in every frame of video (taking into account both the camera motion, and object motion), then statistically "weight" the data in regards to statistical accuracy, and then reconstruct each frame with the most accurate information available, resulting in better, cleaner detail.
In addition, MotionDSP sophisticated motion estimation algorithm reconstruction of scenes with "complex" motion -- for example videos with moving objects, camrera shake, or even both.
Superior to "Frame Averaging"
Competitive "video clarification" tools use a method called “frame-averaging” or “frame-integration,” which requires hours of manual effort by highly skilled and highly trained forensic experts. Frame averaging starts with applying stabilization, then registering each video frame "by hand,", followed by "frame averaging,", which averages the values of each pixel together,
Ikena's super-resolution has several advantages over frame averaging:
- Better results: Frame averaging treats all data as equal, mixing good data with "bad" data -- noise and compression artifacts. Ikena's super-resolution is able to weight "good" data over than "bad" data, resulting in a far superior reconstruction without noise. Ikena can also handle videos with object motion, camera motion, or both.
- Results in seconds, not hours: Ikena does all these steps automatically (motion estimation, super-resolution reconstruction, deblurring, lighting correction, etc) automatically, and in one click. Results from Ikena rely on mathematical algorithms that are exactly reproduceable, every time.
- Faster to learn and use: Because Ikena automates image processing into a single step and provides simple presets, users (even non-forensic experts) can get good results from Ikena with only a few hours of training.
See Ikena in use below.