Téma:High dimensional image similarity criteria for image registration
Vypsáno jako:
Popis:We shall study image similarity criteria for multichannel (e.g. color) images, as well as to evaluate similarity based on texture and other local image descriptors. The tasks may include:

1. Implementing a high-dimensional nearest neighbor mutual information estimator for the ITK or Elastix frameworks and find suitable parameters and optimizer.

2.Compare the nearest neighbor and kernel-based entropy estimators.

3. Study local dimensionality estimation from sample data.

4. Study estimation of mutual information from quantized and noisy data.

5. Improve the stability and speed of the estimator.
Literatura:1. Jan Kybic. High-dimensional mutual information estimation for image registration. In ICIP 2004/ 2. Kybic Jan and Vnučko Ivan. "Approximate All Nearest Neighbor Search for High Dimensional Entropy Estimation for Image Registration." Signal Processing, vol. 92, no. 5, pp. 1302-1316, 2012. 3. Kybic Jan. "High-Dimensional Entropy Estimation for Finite Accuracy Data: R-NN Entropy Estimator." IPMI2007: Information Processing in Medical Imaging, 20th International Conference, pp. 569-580, July 2007.
Realizace:SW projekt
Vypsáno dne:29.05.2018