Distinctive Image Features from Scale-Invariant Keypoints
International Journal of Computer Vision
| Hi-index | 0.00 |
Among the tumor suppressors identified in Drosophila, lgl is one of a few whose proliferative phenotype is shown to be secondary to a loss of cell polarity. Studies of lgl mutant phenotypes are likely to contribute to our understanding of both tumorigenesis as well as neural development mechanisms. However, alterations of neuronal development as a result of key protein mutations are not easy to describe without quantifiable parameters. This work presents a fully automated imaging technique that involves skeletonization and histogram-based features such as kurtosis to find and quantify discriminative morphological phenotypes that can be used to automatically distinguish normal wild-type neurons from lgl mutant neurons.