Supervise Less, See More: Training-free Nuclear Instance Segmentation with Prototype-Guided Prompting

Meta has launched SAM 3, the latest iteration of its Segment Anything Model, which significantly enhances segmentation capabilities by improving accuracy, boundary quality, and robustness in real-world scenarios. This update is the most substantial since the model's inception, aiming to provide more reliable segmentation for both research and production workflows.

A new framework called SAM-MI has been introduced to enhance open-vocabulary semantic segmentation (OVSS) by effectively integrating the Segment Anything Model (SAM) with OVSS models. This framework addresses challenges such as SAM's tendency to over-segment and the difficulties in combining fixed masks with labels, utilizing a Text-guided Sparse Point Prompter for faster mask generation and Shallow Mask Aggregation to reduce over-segmentation.

Image Diffusion Models have demonstrated emergent temporal propagation capabilities in videos, showcasing their potential to enhance video generation and editing processes. This development highlights the growing sophistication of AI technologies in visual media.

The recent introduction of SCALER, a collaborative framework for label-deficient concealed object segmentation (LDCOS), aims to enhance segmentation performance by integrating consistency constraints with the Segment Anything Model (SAM). This innovative approach operates in alternating phases, optimizing a mean-teacher segmenter alongside a learnable SAM to improve segmentation outcomes.

A new framework called Granular Computing-driven SAM (Grc-SAM) has been introduced to enhance prompt-free image segmentation, addressing limitations in the existing Segmentation Anything Model (SAM). Grc-SAM employs a coarse-to-fine approach, improving foreground localization and enabling high-resolution segmentation through adaptive feature extraction and fine patch partitioning.

MedSAM-3 has been introduced as a text promptable medical segmentation model designed to enhance medical image and video segmentation by allowing precise targeting of anatomical structures through open-vocabulary text descriptions. This model builds on the Segment Anything Model (SAM) 3 architecture, addressing the limitations of existing methods that require extensive manual annotation for clinical applications.

The introduction of DEAP-3DSAM, or Decoder Enhanced and Auto Prompt SAM, marks a significant advancement in 3D medical image segmentation, building on the capabilities of the Segment Anything Model (SAM). This new model addresses limitations in spatial feature retention and the reliance on manual prompts, which have hindered previous attempts at applying SAM to 3D images.

A new dataset named CellFMCount has been introduced, consisting of 3,023 images from immunocytochemistry experiments, which includes over 430,000 manually annotated cell locations. This dataset aims to address the challenges of accurate cell counting in biomedical research, particularly in cancer diagnosis and immunology, where traditional manual counting methods are labor-intensive and prone to errors.