Classification of Stained Embryonic Images of Drosophila

Gene expression images are of great interest to the biology community, whose databases are growing by leaps and bounds. Indexing and retrieval of these images become extremely necessary as the size of the databases increase. We consider one such problem here – given an embryonic image stained for gene expression, can we classify them into stages corresponding to the developmental morphology of Drosophila? For this purpose, several preprocessing steps – rotation correction, classification into dorsal and lateral views, and alignment of anterior and posterior ends were performed. Finally the images were classified into morphologically similar groups (in terms of embryo development). Different shape based features were extracted and used for classification. Details of the proposed approach and the results of using this approach on a pilot dataset are presented in this paper.