PROJECT TITLE :
Pairwise Latent Semantic Association for Similarity Computation in Medical Imaging
Retrieving medical images that gift similar diseases is a full of life analysis space for diagnostics and therapy. But, it can be problematic given the visual variations between anatomical structures. In this paper, we propose a brand new feature extraction methodology for similarity computation in medical imaging. Instead of the low-level visual appearance, we tend to style a CCA-PairLDA feature representation technique to capture the similarity between pictures with high-level semantics. 1st, we have a tendency to extract the PairLDA topics to represent a picture as a mix of latent semantic topics in a picture pair context. Second, we have a tendency to generate a CCA-correlation model to represent the semantic association between a picture pair for similarity computation. Whereas PairLDA adjusts the latent topics for all image pairs, CCA-correlation helps to associate a private image pair. During this approach, the semantic descriptions of a picture pair are closely correlated, and naturally correspond to similarity computation between pictures. We evaluated our technique on 2 public medical imaging datasets for image retrieval and showed improved performance.
Did you like this research project?
To get this research project Guidelines, Training and Code... Click Here