Grounded-SAM
PulseAugur coverage of Grounded-SAM — every cluster mentioning Grounded-SAM across labs, papers, and developer communities, ranked by signal.
-
New method enhances zero-shot segmentation for UAV tower inspections
Researchers have developed a new method called Saliency-Depth Conditioning to improve zero-shot segmentation of communication-tower components in cluttered UAV imagery. This approach combines visual saliency with monocu…
-
OptiSight framework integrates semantic reasoning and geometric control for embodied navigation
Researchers have introduced OptiSight, a novel framework designed to enhance autonomous indoor navigation by integrating semantic reasoning with geometric control. This system utilizes a vision-language model (VLM) with…
-
New methods enhance 3D scene understanding for autonomous vehicles · 2 sources tracked
Two new research papers introduce novel methods for semantic occupancy estimation in autonomous driving, aiming to improve 3D scene understanding. The first, Easy3D-Labels, generates 3D pseudo-ground-truth labels using …
-
New SAM-MI framework boosts open-vocabulary semantic segmentation
Researchers have developed SAM-MI, a novel framework designed to enhance open-vocabulary semantic segmentation (OVSS) by integrating the Segment Anything Model (SAM). This framework addresses SAM's tendency to over-segm…
-
ZODS-RS pipeline offers zero-training detection and segmentation for remote sensing
Researchers have developed ZODS-RS, a novel pipeline designed for zero-training object detection and segmentation in remote sensing imagery. This system integrates dense features from DINOv3 with SAM-style proposals to …
-
Peking University's Imagine2Act enables robots to 'imagine then act' for household tasks
Researchers from Peking University have developed Imagine2Act, a novel framework enabling robots to perform intricate household tasks with high precision. The system first "imagines" the desired outcome by generating a …