SegTHOR: Segmentation of THoracic Organs at Risk in CT images
Vincent Lu
Using Image Segementation Techniques to delinearate target tumors and healthy organs located near the target tumor
Radiation therapy is the common choice for lung and esophageal cancer. To begin therapy, irradiation planning is required which involved the manual delineation of tumor and healthy organs which are known as Organs at Risk (OAR). Currently, this delineation is done manually which is tedious and the source of reproducibility errors as organs can vary between patients and can be difficult to spot in CT images.
The purpose of SegTHOR is to automate this process with high levels of accuracy to improve this process. The organs of focus are heart, aorta, trachea, esophagus.
Page Views:
Automating the delineation of organs at risk will improve the process of radiation therapy and the biomedical field. SegTHOR was initially a challenge at IEEE ISBI 2019. Since then, many different variations of AI models have been attempted to improve the image segmentation model.
