A NEURAL NETWORK APPROACH FOR BLOCK CODING IMAGE COMPRESSION | ||||
The International Conference on Electrical Engineering | ||||
Article 30, Volume 2, 2nd International Conference on Electrical Engineering ICEENG 1999, November 1999, Page 278-290 PDF (2.51 MB) | ||||
Document Type: Original Article | ||||
DOI: 10.21608/iceeng.1999.62510 | ||||
View on SCiNiTO | ||||
Author | ||||
M. SHAARAWY IBRAHIM | ||||
Associate Professor, Dpt. of Computer & OR, Military Technical College, Cairo, Egypt. | ||||
Abstract | ||||
This paper presents a scheme for image compression using block coding by vector quantization technique. This scheme achieves promising results in compression ratio and image quality. Although the encoding time represents a big problem in such solutions, it is improved using a self-organizing feature map (SOFM) neural network. Feature extraction reduces the dimensionality of the problem and enables the neural network to be trained on an image separate from that for testing. Although the time complexity has been reduced, the image quality is also affected by a slight value, which can be accepted in many situations. | ||||
Keywords | ||||
Block Coding; Vector Quantization; Self-Organizing Feature Map; Moment Invariant; Feature Extraction | ||||
Statistics Article View: 87 PDF Download: 175 |
||||