Segmentation Mask
Segmentation masks are digital representations of object boundaries within images, crucial for various computer vision tasks. Current research focuses on improving the accuracy and efficiency of generating these masks, particularly in low-data regimes, exploring methods like data augmentation, model re-adaptation, and the utilization of foundation models such as SAM (Segment Anything Model) and diffusion models. These advancements are significantly impacting fields like medical imaging, autonomous driving, and agricultural technology by enabling automated analysis and improved decision-making in data-scarce or complex scenarios.
Papers
Importance of Aligning Training Strategy with Evaluation for Diffusion Models in 3D Multiclass Segmentation
Yunguan Fu, Yiwen Li, Shaheer U. Saeed, Matthew J. Clarkson, Yipeng Hu
Score-Based Generative Models for Medical Image Segmentation using Signed Distance Functions
Lea Bogensperger, Dominik Narnhofer, Filip Ilic, Thomas Pock