STUDIES ON OBJECT RECOGNITION FROM DEGRADED IMAGES USING NEURAL NETWORKS

The objective of this paper is to study the performance of artificial neural network models for recognition of objects from poorly resolved, noisy, and transformed (scaled, rotated translated) images, such as i mages reconstructed from sparse and noisy data in a sensor array imagi ng context. Noise and sparsity of data in the imaging context result i n degradation of quality of the reconstructed image as a whole, instea d of affecting it in the form of local corruption of the image pixel i nformation as in many image processing situations. Hence, (i) neighbou rhood processing methods for noise cleaning may not be suitable, (ii)f eature extraction cannot be reliably performed and (iii) model-based m ethods for classification cannot easily be applied. In this paper, we show that neural network models can be used to overcome some of the di fficulties in dealing with degraded images as obtained in an imaging c ontext.