An automatic framework for standardization of drosophila embryonic images

Embryonic images of Drosophila melanogaster were introduced as an important tool for the discovery of gene-gene interaction, which can be characterized as a computational problem of matching expression patterns of embryos. As the first step of the computational system, image standardization aims to build pixel-to-pixel correspondence so that the comparison between embryonic images are biologically meaningful. Standardization of embryonic images is challenging due to substantial variations of imaging conditions and the appearances of embryos. In this paper, we propose a framework for automatic standardization of embryonic images. Experiment results show the significant improvement achieved by the proposed approach.