Adaptive Compression Algorithm For Full Colour Three Dimensional Integral Images

The aim of the work is to compress digitised images, which contain three-dimensional information, in order that they can be efficiently stored and communicated. The nature of the three-dimensional images investigated is such that there is strong correlation between sub-image elements. It is proposed that to exploit this coiTelation effect the compression technique is based around the three dimensional Discrete Cosine Transform. To enable accurate compression and de-compression of image data, the scanned image must be aligned coiTcctly to take advantage of the inherent structure imposed by the image capture system. A novel method of eiTor angle measurement is proposed which is based around the Hough Transform. This method provides an automated method of measurement of misalignment. An adaptive algorithm based upon the Discrete Cosine Transform is proposed, this algorithm utilises a dynamic mapping process, which arranges the data in such a way that the correlation effects are exploited efficiently. Simulation studies are presented which show that the proposed adaptive algorithm is superior to existing techniques using a number of performance measures. The adaptive technique is extended and applied to full colour images. A series of tests using various filters are earned out to determine which filter type is best suited for the chrominance bandwidth. A series of Hosaka plots are used to compare the performance of the various filter configurations. Simulation studies clearly show that the adaptive algorithm is superior to baseline JPEG for the compression of full colour Integral Images. The application of compression techniques to images containing full parallax information in both horizontal and vertical directions is demonstrated. In order to apply compression techniques to the type of images investigated the pixel structure of the sub-images has to be manipulated. A series of extensions to this work are proposed which, if implemented, would allow compression and communication of moving images hence enhancing the probability of three-dimensional television being commercially viable.

Name dc:type.qualificationname

PhD

Level dc:type.qualificationlevel

Doctoral

Grantor dc:publisher.institution

De Montfort University

Year dc:date.issued

2001

Author dc:creator

• Zaharia, Ramona

Licence

• https://dora.dmu.ac.uk/bitstreams/0a84b4d7-742e-4ec2-aa26-2e473ecb9df1/download