Detection and Localization of Drosophila Egg Chambers in Microscopy Images

Drosophila melanogaster is a well-known model organism that can be used for studying oogenesis (egg chamber development) including gene expression patterns. Standard analysis methods require manual segmentation of individual egg chambers, which is a difficult and time-consuming task. We present an image processing pipeline to detect and localize Drosophila egg chambers that consists of the following steps: 1. superpixel-based image segmentation into relevant tissue classes; 2. detection of egg center candidates using label histograms and ray features; 3. clustering of center candidates and; 4. area-based max. likelihood ellipse model fitting. Our proposal is able to detect 96% of human-expert annotated egg chambers at relevant developmental stages with less than 1% false-positive rate and improves the mean adjusted Rand score (ARS) from 0.75 using common watershed technique to 0.86, which is adequate for the further analysis. Experimental results Input images (Top), initial segmentation (middle) followed by the detected centers (cluster means) as dots and the fitted ellipses in green (bottom). Expert drawn bounding boxes are shown as red rectangles (not all eggs are annotated). Further examples below.