Intelligent system to restore colour images
Intelligent system to restore colour images
Contaminated digital images may become clearer, thanks to a totally new intelligent system for colour image restoration developed at La Trobe University.
The new intelligent system is designed to reduce disturbance or 'noise' which may blur or distort an image, resulting in a clearer picture. It gives clients real time access to a web site which 'filters' out the 'noise'.
For example, a worn old photograph can be scanned and transmitted as a Jpg file. But often markings or faded areas on the old print result in disturbances or 'noise' in transmission. In other cases, a good image can arrive at its destination after being distorted in transmission.
The receiver of such distorted images can immediately divert the file to a web site which cleans up the image.
The technique can be applied commercially in a number of areas, as well as military. For example, military targets-often identified from old photographs-can be clarified, enabling the target to be identified more clearly.
Dr Dianhui Wang, a lecturer in La Trobe University's School of Computer Science and Computer Engineering, has produced a prototype of the new system. He is currently working to bring it to a stage where an industry partner could take steps towards commercialisation.
Dr Wang says that a degraded image may be caused by various factors such as atmospheric turbulence, distortions in the optical imaging system, lack of focus, sensor or transmission noise injection, coding techniques, and object or camera motion.
"The task of image restoration is to remove these degradations to enhance the quality of the image for further use in domain applications. Image restoration can be defined as a problem of estimating a source image from its degraded version.
"In the past, considerable studies using various approaches have been investigated to solve this fundamental and important issue for image processing.
"Our system is a pattern learning based image restoration technique using neural networks to enhance the quality of images, where a prior knowledge of the image dependent edge information was incorporated into the regularised error measure to improve the upper bound estimation of the high frequency content."
Related Article: