Image abstraction is a method of enhancing or high lighting the visualization of main information contents of image and smoothing the other portion. It is quite important for many multimedia applications like cartooning and animation where symbolic presentation matters; hence it is an important part in the field of image processing. In this Paper we study a neural network based approach & also the other methods of Image Enhancement. A neural network based approach which not only abstract the image but also avoids the enhancement of noise. The algorithm uses the image abstraction technique for detecting the information density in different parts of image then accordingly operates the smoothing filter and after filtering the information’s of edges are recombined with the filtered image. The technique also utilizes the Neural Network for filtering noise generated edge patterns. Hence the approach not only enhances the image but also avoids the enhancement of noise. After comparison with other enhancement techniques we conclude that it improves the perception; remove noise while maintaining the structure information intact it is also found that the Neural Network technique is quite fast.

Full text: IJECCE_585_Final.pdf