Project Motivation:
Introduction:
Our group consists of two electrical engineers, who understand the complexities and tradeoffs of working with applications that take up a lot of storage memory and bandwidth. The possibility of sending "thumbnail" like pictures through the internet and then blowing them up at the receiving end without a perceptible difference in the quality of the images fascinated us. We looked into various methods of implementing a reliable and efficient image scaling algorithm.
Many of the algorithms worked well, but we were still left with blurring in the images. We explored various edge-detection algorithms. A common solution was the use of high-pass filters to enhance edges. These result in definite improvement in the image quality, but also have the tendency to amplify whatever noise is already present in the original image. We then explored noise reduction to solve this issue and we had to deal with the tradeoff between accuracy of edge detection and noise reduction.
We were directed by Professor Brian Wandell to look into bilateral filtering as a possible solution and as we found out it was indeed effective. We are now able to enhance images using our image scaling algorithm with bilateral filtering and have found these images to look the best when compared to other methods.
Future Scope:
The future for this kind of system is the ability to compress and store images and then expand them when needed at the receiving end with minimal delay. There are a lot of areas where this kind of system can be used. We have identified remote management as one of the most potential applications are eager to see how far this can go.