The Graz-based microscopy image analysis software company KML Vision recently published an innovative computer-assisted method to explore the microstructure of cardiac tissue. The AI-backed approach relies on deep learning techniques to detect tiny morphological structures of cardiac cells in microscopy images. This groundbreaking method allows life science researchers to train their own deep learning algorithms to examine cardiac cell structures without requiring any coding skills.   

Cardiovascular disease is the most common cause of fatality in developed countries according to WHO statistics. The onset of cardiovascular disease is often associated with morphological changes on a cellular level. Some of those processes are increased lipid synthesis and reduced mitochondrial energy production. Understanding these pathological processes better can help develop new treatments strategies and improve patient survival rates.

KML Vision’s deep learning-based method enables the fast and reliable detection and quantification of cellular structures such as mitochondria and lipid droplets in cardiac tissue. The algorithm training itself requires minimal input from researchers and is performed on an intuitive and user-friendly graphic interface. A small number of representative cardiac tissue images with respective annotations of the objects of interest needs to be provided as training input. No coding or programming intervention is required from the researcher.

The algorithm actively learns from the input data and is able to automatically detect and measure cardiac cell structures and properties even in particularly large datasets containing thousands of images. The proposed method yields an excellent performance in terms of accuracy and saves researchers a substantial amount of time and manual effort.

Learn about this innovative myocardium tissue segmentation technique in detail in KML Vision’s case study: https://www.kmlvision.com/resources/case-studies/deep-learning-tem-myocardium-assay/