MIT researchers have developed MultiverSeg, an interactive AI tool enabling efficient, user-driven segmentation of biomedical image datasets without prior model training.
Key Details
- 1MultiverSeg allows users to annotate images through clicks and scribbles, reducing manual input over time.
- 2The system does not require presegmented data or machine learning expertise for new tasks.
- 3By the ninth image, only two user interactions are needed for accurate segmentation, outperforming existing tools.
- 4Applicable across imaging types such as X-ray and adaptable to a range of biomedical image datasets.
- 5Supported by Quanta Computer and the NIH, and benchmarked against state-of-the-art segmentation tools.
Why It Matters
Manual image annotation is a major bottleneck in radiology and clinical research. This tool could significantly expedite dataset preparation for studies and AI development, lowering expertise and resource barriers and potentially accelerating clinical translation.
Source
EurekAlert
Related News
•EurekAlert
AI-Powered OCT Enables Rapid 'Optical Biopsy' for Early Endometrial Cancer Detection
A team at Washington University has developed a catheter-based 3D OCT system with AI to quickly and noninvasively detect early endometrial cancers.
•EurekAlert
AI Clinical Reasoning in Diagnostics and Digital Fatigue in Healthcare
Recent JMIR features explore large language models in clinical diagnostics and digital fatigue among healthcare professionals.
•EurekAlert
KAIST, MIT, Microsoft Develop Efficient AI Image Upsampling for Robotics
KAIST, MIT, and Microsoft have created 'Upsample Anything,' a training-free AI method to restore high-resolution visual data from compressed images with up to 16x improved GPU memory efficiency.