Computer assisted assessment of histopathology images

Nerve damage occurs in various ways such as accidents, traumas, surgical operations and inflammatory diseases. Damages in the nerve cells affect muscle cells that are innervated by the nerves. Directly observing the nerve damage may not be possible in all circumstances. To observe the effect of the nerve damage, muscle cells that are connected to the nerve cells need to be histomorphologically evaluated. In treatment of nerve damages, nerve regeneration is aimed to be enhanced by placing tissues such as nerve auto graft (replacing the cut nerve to its original place), vessel, vessel + cartilage, cartilage in place of the damaged area. However, regeneration in nerve cells is known to be very minimal. The success of regeneration is measured using microscopic analysis of histopathological tissue samples. However, this approach is slow, error prone, irreproducible and also depends on the expert's personal assessment. This study presents a Computer Aided Diagnosis (CAD) system that enables quantitative assessment of muscle cell tissues that are directly affected by a nerve damage. Proposed approach aims to reduce subjectivity in assessment of various regeneration alternatives by performing morphological operations on images and providing quantitative results of muscle cell nuclei counts and density, the ratio of nucleus' parallelism to the other tissue components, the length of cell nuclei, and the color density of nuclei staining. This study is carried on 530 images of muscle cell tissues containing four damaged and one healthy control group acquired from Wistar albino rats.