accepted short papers

SBN/EAN: 978-90-816079-1-9
Titel: Supplementary Proceedings of Pattern Recognition in Bioinformatics
Subtitel: 5th IAPR International Conference, PRIB 2010
Nijmegen, The Netherlands, September 22-24, 2010

  • Towards Metadata Patterns in Image Databases
    Joris Slob, Amalia Kallergi, Julia Dmitrieva and Fons Verbeek

  • Abstract (short paper)
    Pattern recognition in image databases has focused on homogeneous images from single or limited number of modalities. Biological problems require a more integrated approach that crosses scales and modalities. We present a database architecture that focuses on image metadata enriched with ontology term annotations. A use-case within the research field of Alzheimer’s disease shows how such a database architecture can be a valuable source of investigation and exploration.

  • Efficient Remote Homology Detection
    Antolin Janssen, Evgeni Tsivtsivadze, Jorma Boberg, Tjeerd Dijkstra and Tom Heskes

  • Abstract (short paper)
    We propose an efficient multi-class classification algorithm for remote homology detection (RHD). Unlike methods that treat RHD as a set of binary classification tasks, our algorithm solves a single multi-class classification problem by incorporating information about class-wise correlations among the proteins using joint kernel functions. Furthermore, the proposed method leads to notable reduction in computational time compared to binary classification algorithms. We evaluate our method on the Structural Classification of Proteins database and show that performance is better or comparable to several state-of-the-art algorithms for protein classification and remote homology detection.

  • RelaxGrep: Approximate Graph Searching by Query Relaxation
    Vincenzo Bonnici, Rosalba Giugno, Alfredo Pulvirenti, Alfredo Ferro and Dennis Shasha

  • Abstract (short paper)
    Approximate graph searching plays a key role from biology to computer graphics. Here we propose RelaxGrep, a tool that guides the user to find subgraphs in the database that are isomorphic to a “relaxation” of an original query Q, (i.e., a graph obtained after removing edges or nodes from Q). This is done through a modification of the filtering and verification phase of GraphGrep. Experiments yield promising results on molecular databases.

  • Concurrent Visualization of and Mapping between 2D and 3D Medical Images for Disease Pattern Analysis
    Mei Xiao, Jung Soh, Thao Do, Oscar Meruvia-Pastor and Christoph Sensen

  • Abstract (short paper)
    Brain lesions caused by several neurologic disorders have a range of shapes, sizes and locations. Discovering the patterns of lesion development in patients will assist in evaluating the damages and monitoring the recovery of brain tissues. We have developed a software tool to highlight and display regions of interest in 2D medical images and their 3D mesh model counterparts. This tool can help researchers visualize and compare the brain lesions and tissues in multiple dimensions and various user-controlled configurations.

Machine Learning Group