Medical imaging is the process of obtaining pictorial representations of the inner parts of a human body for clinical examination and medical diagnosis. These images needed to store and transmit for telemedicine applications. Hence, for error free storage and hazel free transmission, the size of the images are reduced using different compression techniques before transmission. For medical image compression, two techniques i.e. Feed Forward Back Propagation Artificial Neural Network (FFBPANN) with different training algorithms and Singular Value Decomposition (SVD) are used to compress and reconstruct them with good quality. To get much better compression with considerable image quality, a hybrid technique which combines FFBPANN with SVD is proposed in this paper. The compressed output is then encoded with Huffman encoding to get compressed bits. This proposed technique of SVD-FFBPANN using LM algorithm provides low Mean Squared Error (MSE), high Peak Signal to Noise Ratio (PSNR) compared to existing algorithms.
Volume 11 | Issue 6
Pages: 225-235