Medical AI : assessing stroke severity | |
| Author | Noppawee Teeraratchanon |
| Call Number | AIT ISPR DSAI no.25-02 |
| Subject(s) | Artificial intelligence--Medical applications Medical informatics |
| Note | An internship report submitted in partial fulfillment of the requirements for the degree of Master of Engineering in Data Science and Artificial Intelligence |
| Publisher | Asian Institute of Technology |
| Abstract | This project was conducted during my five-month internship at AI Brain Lab, from February to June. It focuses on the development of medical image processing software designed to assist in stroke assessment using non-contrast brain CT scans. The system offers two selectable approaches:an automatic brain layer selection approach, which automatically identifies slices corresponding to the Basal Ganglia and Corona Radiata layers, and a manual brain layer selection approach, where users manually select a slice, and the system aligns it to a predefined anatomical template using both rigid and nonrigid registration techniques.In both approaches, the software segments the brain into 10 anatomical regions, calculates the ASPECTS score, and extracts regional Hounsfield Unit (HU) values to support stroke severity evaluation.To ensure accurate and efficient alignment, various configurations of transforms, similarity metrics, optimizers, and interpolators were evaluated. The optimal parameters for rigid registration were selected based on the Structural Similarity Index Measure (SSIM) and computation time, while the non-rigid registration setup was chosen based on the quality of automatic layer selection and processing efficiency.Thesystem wastested on 12 brain CTsamples andevaluated using statistical hypothesis testing. Results demonstrated that the automatic process completed in under three minutes per brain on average, while the manual process required less than one minute per layer on average. The system consistently produced reasonable ASPECTS scores, even in challenging anatomical cases, highlighting its potential as a decision-support tool for radiologists in clinical stroke assessment. |
| Year | 2025 |
| Type | Internship Report |
| School | School of Engineering and Technology |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Data Science and Artificial Intelligence (DSAI) |
| Chairperson(s) | Chaklam Silpasuwanchai |
| Examination Committee(s) | Attaphongse Taparugssanagorn;Chantri Polprasert |
| Scholarship Donor(s) | AIT President;AIT Scholarship |
| Degree | Internship Report (M. Eng.) - Asian Institute of Technology, 2025 |