Sam3
PulseAugur coverage of Sam3 — every cluster mentioning Sam3 across labs, papers, and developer communities, ranked by signal.
- 2026-06-25 research_milestone Researchers introduced the Mask to Concept (M2C) framework to adapt SAM3 for few-shot medical image annotation. source
11 day(s) with sentiment data
-
New synthetic dataset WireSeg-32K improves wire instance segmentation
Researchers have introduced WireSeg-32K, a synthetic dataset designed to improve wire instance segmentation. This dataset comprises 32,000 images with instance masks and depth maps, generated using a co-simulation pipel…
-
SAM3-LoRA adapts foundation model for defect segmentation using parameter-efficient technique
Researchers have developed SAM3-LoRA, a parameter-efficient adaptation technique for the SAM3 foundation model, specifically for multi-class structural defect segmentation. This method utilizes Low-Rank Adaptation (LoRA…
-
OPUS framework simplifies open-vocabulary detection with strong performance
Researchers have introduced OPUS, a novel unified framework for open-vocabulary detection designed for simplicity and effectiveness. Unlike previous complex systems, OPUS leverages semantic-rich visual representations a…
-
New ComfyUI nodes enable INT8 quantization for SAM3 models, saving storage space
A developer has created ComfyUI nodes that support ConvRot INT8 quantization for the SAM3 and SAM3.1 models. This optimization significantly reduces storage space and VRAM usage by keeping model weights in 8-bit precisi…
-
New framework AffectOmni enhances LLM affective reasoning with verifiable human cues
Researchers have developed AffectOmni, a framework designed to improve affective reasoning in multimodal large language models (MLLMs). This new system focuses on people-centric cues like micro expressions and body lang…
-
New method GAP-SAM improves AI-generated image manipulation localization
Researchers have developed GAP-SAM, a novel method for localizing manipulations in AI-generated images. This approach addresses limitations in existing techniques, such as poor out-of-distribution performance and a tend…
-
SAM3 and Gemini-3.1 Pro achieve third place in LSVOS Challenge
Researchers have developed a novel two-stage approach for video object segmentation, achieving third place in the 8th LSVOS Challenge's MeViS-Text track. The method first utilizes Gemini-3.1 Pro to break down video even…
-
New RISE framework integrates 3D tracking and vision-language reasoning for roadside analysis
Researchers have introduced RISE (Roadside Infrastructure Sequence Understanding and Evaluation), a novel framework designed for analyzing roadside traffic sequences. This framework integrates metric 3D tracking using i…
-
VOS-Agent framework takes 1st place in LSVOS Challenge
Researchers have developed VOS-Agent, a novel framework for complex video object segmentation that improves upon existing methods like SAM3. VOS-Agent utilizes a Target Perception and Routing Agent to categorize targets…
-
New benchmark H2R-Bench reveals limitations in human-to-robot video generation
Researchers have introduced H2R-Bench, a new benchmark designed to evaluate video generation models' ability to translate human manipulation videos into robot-centric demonstrations. The benchmark addresses the challeng…
-
SAM3 model uses spatial guidance for better PV segmentation in remote sensing
A new research paper evaluates the effectiveness of different prompting strategies for segmenting small-scale photovoltaic (PV) systems in remote sensing imagery using the SAM3 vision-language foundation model. The stud…
-
New SAM3-based method enables annotation-free surgical instrument segmentation
Researchers have developed a novel two-stage framework for segmenting surgical instruments in endoscopic images without requiring manual annotation. This approach utilizes the Segment Anything Model 3 (SAM3) with a gene…
-
New Visual Token Codec Boosts ViT Feature Compression Efficiency
Researchers have developed a new method called the Visual Token Codec (VTC) to compress intermediate features in Vision Transformer (ViT) models. VTC effectively utilizes the spatial correlations inherent in ViT patch t…
-
SC-Diff model enhances visible-to-infrared image translation with semantic calibration
Researchers have developed SC-Diff, a novel diffusion model designed for translating visible images into infrared representations. This framework enhances semantic consistency by calibrating self-attention mechanisms wi…
-
New adversarial attack targets SAM3 image segmentation models
Researchers have developed Universal Concept Disruption (UCD), a novel adversarial attack specifically designed to target SAM3 image segmentation models. UCD learns a single image perturbation that can disrupt the model…
-
Tarot-SAM3 framework enhances SAM3 for any referring expression segmentation
Researchers have developed Tarot-SAM3, a new framework designed to improve referring expression segmentation (RES) by enabling the Segment Anything Model 3 (SAM3) to handle any natural language expression. The framework…
-
New prompt learning method boosts medical image segmentation accuracy
Researchers have developed Few-Shot Concept Prompt Learning (FS-CPL) to improve the performance of segmentation foundation models like SAM3 and Medical SAM3 in medical imaging. This new method learns a continuous concep…
-
New frameworks enhance multimodal visual tracking accuracy and efficiency
Two new research papers propose novel frameworks for unified multimodal visual tracking, aiming to improve accuracy and efficiency. The first paper introduces ACTrack, an agentic coordination framework that treats vario…
-
SAM3 enhances remote sensing image segmentation without training
Researchers have developed a novel training-free framework for few-shot segmentation of remote sensing images, leveraging SAM3's capabilities. This approach utilizes SAM3's geometric priors to create category-agnostic e…
-
UltraSAM3: New Foundation Model for Universal Ultrasound Image Segmentation
Researchers have developed UltraSAM3, a novel concept-driven foundation model designed for universal ultrasound image segmentation. This model adapts SAM3 to work with ultrasound-specific image-mask-concept triplets, en…