Abstract #0623

Enhancing organ segmentation performance in foundation models via ensemble learning

Qing Li¹, Yizhe Zhang², Yan Li³, Yajing Zhang⁴, Longyu Sun⁵, Mengting Sun⁵, Qirong Li⁶, Zian Wang⁶, Meng Liu⁵, Xumei Hu⁵, Shuo Wang⁷, and Chengyan Wang⁵

¹Fudan University, Shanghai, China, ²School of Computer Science and Engineering, Nanjing university of science and technology, Nanjing, China, ³School of Medicine, Shanghai Jiao Tong University, Shanghai, China, ⁴GE Healthcare, Beijing, China, ⁵Human Phenome Institute, Fudan University, Shanghai, China, ⁶School of Computer Science, Fudan University, Shanghai, China, ⁷Digital Medical Research Center, School of Basic Medical Sciences, Fudan University, Shanghai, China

Synopsis

Keywords: Analysis/Processing, Machine Learning/Artificial Intelligence, Ensemble learning, Organ segmentation, Foundation model, Accuracy, Fairness

Motivation: The application of foundation models in organ segmentation faces numerous challenges related to accuracy and fairness. Ensemble learning combines the strengths of multiple models and shows potentials to enhance segmentation performance, yet has not been studied in foundation models.

Goal(s): This study aims to improve the accuracy and fairness of foundation models across gender, age and BMI using ensemble learning technique.

Approach: The foundation models(TotalSegmentator, SAM, SAM2, MedSAM and MedSAM2) were used for organ segmentation while 5 ensemble methods were used for model improvement.

Results: Foundation models face notable challenges regarding accuracy and fairness. However, employing ensemble learning has effectively enhanced the performance.

Impact: This study integrates the ensemble learning technique for the first time to enhance the performance of foundation models, potentially reducing costs in time and resources. More importantly, it provides an effective approach for improving foundation model performance in future applications.

How to access this content:

For one year after publication, abstracts and videos are only open to registrants of this annual meeting. Registrants should use their existing login information. Non-registrant access can be purchased via the ISMRM E-Library.

After one year, current ISMRM & ISMRT members get free access to both the abstracts and videos. Non-members and non-registrants must purchase access via the ISMRM E-Library.

After two years, the meeting proceedings (abstracts) are opened to the public and require no login information. Videos remain behind password for access by members, registrants and E-Library customers.

Click here for more information on becoming a member.