Segment Anything Model
PulseAugur coverage of Segment Anything Model — every cluster mentioning Segment Anything Model across labs, papers, and developer communities, ranked by signal.
6 day(s) with sentiment data
-
PolSAM enhances Segment Anything Model for PolSAR data analysis
Researchers have developed PolSAM, a novel adaptation of Meta AI's Segment Anything Model (SAM) designed to improve image segmentation for Polarimetric Synthetic Aperture Radar (PolSAR) data. PolSAM integrates domain-sp…
-
UAV computer vision system analyzes cattle feeding patterns
Researchers have developed a computer vision framework using Unmanned Aerial Vehicles (UAVs) to analyze cattle feeding patterns in commercial feedlots. The system employs YOLOv8 and YOLOv11 models for object detection, …
-
New BALD-SAM framework enhances interactive segmentation with active prompting
Researchers have introduced BALD-SAM, a novel framework for interactive segmentation that leverages Bayesian Active Learning by Disagreement (BALD) to strategically select prompts. This method aims to improve the effici…
-
New THG-Omni framework improves AI video reasoning with graph supervision
Researchers have developed THG-Omni, a novel framework designed to enhance video reasoning capabilities in omni-modal AI models. This system uses a human-centric graph-supervision approach, linking participant tracks fr…
-
AI framework enhances breast cancer diagnostics with multi-view mammogram analysis
Researchers have developed a new AI framework called MSMV-Swin to improve breast cancer diagnostics using mammograms. This framework integrates AI into a human-centric workflow, addressing challenges like the need for d…
-
New framework enhances Segment Anything Model for medical image segmentation
Researchers have developed SAM-DCE, a new framework designed to improve the prompt-free segmentation capabilities of the Segment Anything Model (SAM) for medical imaging. This approach addresses SAM's tendency to lose f…
-
New framework generates pseudo-labels for material CT scans without GPUs
Researchers have developed TEMPL, a novel framework for generating pseudo-labels in material CT scans without requiring pre-trained models or GPUs. This model-free approach utilizes thresholding, morphological dilation,…
-
PR-MaGIC framework refines image segmentation prompts without training
Researchers have developed PR-MaGIC, a novel framework designed to improve in-context image segmentation by refining prompts. This training-free method utilizes gradient flow from a model's mask decoder to enhance promp…
-
SAM 2.1 fine-tuned for improved building component segmentation in 3D reconstruction
Researchers have fine-tuned SAM 2.1, a vision foundation model, on the HBD dataset to improve the segmentation of building components in 2D images. This adaptation led to a significant increase in mIoU from 49.9% to 69.…
-
New pipeline automates 3D dendrite segmentation in brain imaging
Researchers have developed an automated pipeline for segmenting 3D dendrites in SBF-SEM images, a crucial step for understanding brain plasticity. This system integrates YOLOv6 with the Segment Anything Model (SAM) for …
-
Unsupervised MoSA framework learns object segmentation from motion data
Researchers have developed Motion-Grounded Segment Anything (MoSA), an unsupervised framework designed to overcome the reliance on massive manual annotations that plague models like the Segment Anything Model (SAM). MoS…
-
New attack method targets advanced AI segmentation models
Researchers have developed AdvPCS, a novel adversarial attack method designed to compromise promptable concept segmentation models, including the latest SAM3. This attack technique focuses on universal cross-prompt tran…
-
SAMReg uses Segment Anything Model for advanced image registration
Researchers have developed SAMReg, a novel image registration algorithm that leverages the Segment Anything Model (SAM) for improved accuracy and efficiency. This method utilizes a region-of-interest (ROI) based corresp…
-
New pipeline enhances image analysis for social science research
Researchers have developed a new pipeline called Perceive, Refine, Reason (PRR) designed to improve the measurement of specific objects and their placement within images for social science research. This system integrat…
-
PEFT techniques enhance SAM for liver tumor segmentation
Researchers have explored parameter-efficient fine-tuning (PEFT) techniques for segmenting liver tumors in CT scans using the Segment Anything Model (SAM). The study compared several PEFT methods, including LoRA, QLoRA,…
-
MedSAM-3 enhances medical image segmentation with text prompts and LLM agents
Researchers have introduced MedSAM-3, a new model designed for medical image segmentation that leverages text prompts for precise targeting of anatomical structures. By fine-tuning the Segment Anything Model (SAM) archi…
-
New benchmark and survey advance remote sensing image segmentation
Researchers have introduced VPRef, a new benchmark for referring remote sensing image segmentation designed to address performance degradation caused by visual and textual domain drift. This benchmark, featuring over 46…
-
New spectral adapters enhance SAM for medical image segmentation
Researchers have developed two novel spectral adapters, DiSECT and SiGA, designed to enhance the Segment Anything Model (SAM) for segmenting colorectal liver metastases (CRLM) in CT scans. These adapters aim for paramet…
-
New SARTM framework adapts SAM for RGB-thermal segmentation
Researchers have developed SARTM, a new framework designed to adapt the Segment Anything Model (SAM) for RGB-thermal (RGB-T) semantic segmentation. SARTM fine-tunes SAM with LoRA layers and incorporates language guidanc…
-
AI model adapted for precise riverbank erosion analysis in Bangladesh
Researchers have adapted the Segment Anything Model (SAM) to analyze riverbank erosion in Bangladesh using historical Google Earth imagery. This fine-tuned model, which focused on adapting the mask decoder while keeping…