Here you can see the power of Amped Five in full effect. The first section shows some image filtering, while the second  proposes some multi-step processing on video files related to typical situations.

Image samples

Motion deblurring: pass the mouse over the images to correct blur caused by moving objects.

Optical deblurring: pass the mouse over the image to correct blur caused by wrong focus.

Contrast enhancement: pass the mouse over the images to see what's hidden in the dark. 

Smart zoom: pass the mouse over the image to have a better detail resolution (4x zoom).

Denoise: pass the mouse over the images to reduce noise. 

Deblocking: pass the mouse over the images to remove jpeg compression artifacts. 

Undistort: pass the mouse over the images to correct geometric distortions caused by wide-angle optics.

Spectrum filtering: pass the mouse over the image to remove the interferences.

Perspective correction: see images from a different angle or compensate perspective effect.

Unroll: convert images from omnidirectional (dome) to panoramic.

Video samples 

The following samples show some more complex processing on videos. In order to better understand the process, each step of the sequence is shown.

Extracting a license plate number from an extremely dark video

Step 1. This is the original video, almost totally dark. You can see some small glare at the end of the video.

Step 2.  When correcting the intensity curve, a very strong noise is shown.

Step 3.  We can identify the area where the license plate is and stabilize it.

Step 4. This is the final result obtained with frame integration. Not bad, considering the starting point: from totally black frames!

Stabilizing a video

Step 1. This is the original video. As you can see, the image is strongly shaking. 

Step 2. This is the result obtained by stabilizing the original video.

Enhancing a label in a low light environment

Step 1. This is the original video. There is quite much noise and the contrast is rather weak.

Step 2. Here we stabilize the image according to the label.

Step 3. This is the result after the integration of the stabilized video... but more can be done...

Step 4. Applying contrast enhancement allows to clearly read the numbers on the label!

Enhancing a license plate taken from different points of view

Step 1. This is the original video, taken by walking all around a scooter.

Step 2. Here only some frames are selected.

Step 3. We correct the perspective in order to simulate the license plate view from the same angle in different frames.

Step 4. Then we enhance brightness and contrast.

Step 5. This is the final result. Even in this case the license plate becomes readable.

Visualizing a scene of interest from an interlaced and multiplexed video

Step 1.  This is the original file (actually cropped to show only the lower right quarter). As you can see, different scenes are interlaced together, in the even and odd fields of the image.

Step 2.  The first step is to separate the fields, in this way the number of frames is doubled, while the height of each frame is halved.

Step 3.  Now we can reconstruct the missing lines with interpolation. Separation and interpolation have been separated in two steps just to make it clearer; in Five the two operations are usually done at the same time.

Step 4.  Now we can select only the frames of interest in the sequence and visualize them more easily.