Multi-resolution border segmentation for measuring spatial heterogeneity of mixed population biofilm bacteria.

Multi-resolution image clustering and segmentation interactive system has been developed to analyze the interaction between clusters of heterogeneous microbial populations residing in biofilms. Biofilms are biological microorganisms attached to surfaces, which develop a complex heterogeneous three-dimensional structure. The hierarchical structural analysis concept underlying multi-resolution image segmentation is that the clusters will be more complex and noisy for higher-resolution while less complex and smoother for lower-resolution image. This hierarchical structure analysis can be used to simplify the image storage and retrieval in well-mixed populations. We are proposing an algorithm that combines Fuzzy C-Means, SOM and LVQ neural networks to segment and identify clusters. The outcome of the image segmentation is quantified by the number of cluster objects of each kind of microorganism within sections of the biofilm, and the centroid distances between the identified cluster objects. Experimental evaluations of the algorithm showed its effectiveness in enumerating cluster objects of bacteria in dual-species biofilms at the substratum and measuring the associated intercellular distances.

High performance bio-image database retrieval using MPI.

Fast and accurate 3D object reconstruction and partial 3D component retrieval from 2D image slices represent a difficult and challenging problem. To group related objects on different layers in an image stack, image segmentation and sequential matching of adjacent 2D objects have to be preformed. Object matching involves heavy computing and is time consuming. In this paper, we propose a new approach for parallel implementation of object contour matching and partial 3D component retrieval based on image contour structure. The method has been implemented in MPI on a SGI Origin 2000 machine. The experimental results show a good speedup for sequential object matching and partial 3D component retrieval.

3D Reconstruction Using Image Contour Data Structure.

Fast and accurate 3D object reconstruction from 2D image slices represents a difficult and challenging problem. Scientists have to either manually define the boundary of the partial object which is time consuming and may lack accuracy or provide the object boundary data which requires all objects be segmented. In this paper, we propose a new method for 3D reconstruction based on optimal image contour mapping. We also propose a novel date structure to represent the corresponding 3D objects. In our approach, all object contours in the same slice as well as adjacent slices are automatically segmented and combined in a hierarchical tree data structure. This data structure allows fast 3D object retrieval and 3D component analysis.

Protein secondary structure prediction using neural network and simulated annealing algorithm.

Predicting the secondary structure of a protein (alpha-helix, beta-sheet, coil) is an important step towards elucidating its three dimensional structure, as well as its function. In this research we use a multilayer feed forward neural network for protein secondary structure prediction. The RS126 data set was used for training and testing the proposed neural network. We combined neural network and simulated annealing (SA) to further improve on the accuracy of protein secondary structure prediction. The results obtained show that by combining the neural network with SA technique improves the prediction accuracy in the range of 2-3%.