IMAGE ANALYSIS FOR ZEBRAFISH VASCULATURE

Zebrafish has become a viable model for various research including vertebrate development, gene expression analysis, human diseases modeling, drug screening and toxicology analysis. Zebrafish have a closed circulatory system, and the mechanisms of vessel formation are highly similar to those in humans. Being able to model the growth of blood vessel in the vasculature system of zebrafish is interesting for understanding both the circulatory system in humans, and for facilitating large scale screening of the influence of various chemicals on vascular development. Zebrafish embryo is an attractive alternative for environmental risk assessment of chemicals since it offers the possibility to perform high throughput analysis in vivo. Intersegmental vessels (ISV) and caudal vein plexus (CVP) undergo active development via angiogenesis. Hence, providing excellent models to study vasculature system. However, the lack of tools for automated analysis of acquired images is a huge bottleneck in utilizing the zebrafish to its full potential.

Most of the current research based on ISV observe the presence or absence of ISVs or perturbation of ISV morphology but do not quantify growth dynamics. Moreover, these analyses are done manually; hence, it is tedious and expensive. All of these factors drive the need for automated image processing methods to quantitatively analyze the imaged embryos. In this work, we have focused on developing image processing algorithms to automatically segment and quantify ISVs of zebrafish embryos that have been exposed to various chemicals. We tested the algorithms on images of zebrafish embryos obtained from screening compounds that may act as an ISV disruptor. The efficiency of segmentation and quantification approach is demonstrated by our experiments of the entire zebrafish vasculature recorded using a fluorescence microscope.

In this work, we have also presented an approach to segment and detect abnormalities in the CVP region of zebrafish embryos due to exposure to chemicals. Morphological changes due to chemicals exposure are modeled based on the proposed gradient weighted co-occurrence histogram of oriented gradients (gCo-HOG). These features are compared to more commonly use gray level co-occurrence matrix (GLCM), histogram of oriented gradients (HOG) features, and co-occurrence histogram of oriented gradients (Co-HOG) features that utilizes distribution of neighboring pixels to capture spatial structure.