A Neural Network based method for Data Compression

In this research we present a method for the compression of the images using Artificial Neural Networks. ANNs trained using a Direct Solution Method and Error Backpropagation are compared. Many steps are taken for the successful compression of the image. The steps proposed are as follows: 1. Image Acqisition, 2. Preprocessing, 3. Segmentation of the image, 4. Preparation of the training pairs. 5. Exclusion of similar training pairs, 6.Compression of image using DSM and EBP and 7. Reconstruction of the image.

Various small and large images such as Digital Mammograms and Lena are compressed and reconstructed.

Data Compression Links

Neural Network Compression Research

Image Compression through Neural Networks

Data Compression Conference

Video Compression using Neural Network and Fuzzy Logic

Data Compression and Traffic Management

References

[1] B. Verma, M. Blumenstein and S. Kulkarni, 1999, A New Compression Technique Using an Artificial Neural Network, Journal of Intelligent Systems, 9, 39-53.
[2] S. Kulkarni, B. Verma, and M. Blumenstein, 1997, Image Compression using a Direct Solution Method based Neural Network, Proceedings of the Tenth Australian Joint Conference on Artificial Intelligence, Perth, Australia. 114-119.
[3] B. Verma, M. Blumenstein, and S. Kulkarni, 1997, A Neural Network Based Technique for Data Compression, Proceedings of the IASTED International Conference on Modelling and Simulation, MSO '97, Singapore, 12-16.