arXiv:2511.13714v1 Announce Type: cross 
Abstract: The Segment Anything Model (SAM) family has become a widely adopted vision foundation model, but its ability to control segmentation granularity remains limited. Users often need to refine results manually - by adding more prompts or selecting from pre-generated masks - to achieve the desired level of detail. This process can be ambiguous, as the same prompt may correspond to several plausible masks, and collecting dense annotations across all granularities is prohibitively expensive, making supervised solutions infeasible. To address this limitation, we introduce UnSAMv2, which enables segment anything at any granularity without human annotations. UnSAMv2 extends the divide-and-conquer strategy of UnSAM by discovering abundant mask-granularity pairs and introducing a novel granularity control embedding that enables precise, continuous control over segmentation scale. Remarkably, with only $6$K unlabeled images and $0.02\%$ additional parameters, UnSAMv2 substantially enhances SAM-2, achieving segment anything at any granularity across interactive, whole-image, and video segmentation tasks. Evaluated on over $11$ benchmarks, UnSAMv2 improves $\text{NoC}_{90}$ (5.69 $\rightarrow$ 4.75), 1-IoU (58.0 $\rightarrow$ 73.1), and $\text{AR}_{1000}$ (49.6 $\rightarrow$ 68.3), showing that small amounts of unlabeled data with a granularity-aware self-supervised learning method can unlock the potential of vision foundation models.

أصبح نموذج Segment Anything (SAM) شائعًا كنموذج أساسي للرؤية، لكنه يواجه صعوبة في التحكم في دقة التقسيم، مما يتطلب غالبًا من المستخدمين إجراء تعديلات يدوية. لمعالجة هذه المشكلة، تم تقديم UnSAMv2، الذي يمكّن من تقسيم أي شيء على أي دقة دون الحاجة إلى تعليقات بشرية. يوسع هذا النموذج استراتيجية التقسيم والفوز الخاصة بسابقه، UnSAM، من خلال اكتشاف أزواج وفيرة من الأقنعة والدقة وإدخال تضمين جديد للتحكم في الدقة يمكّن من التحكم الدقيق والمستمر في مقياس التقسيم. يظهر النموذج فعاليته باستخدام 6000 صورة غير مصنفة فقط.

El modelo Segment Anything (SAM) ha ganado popularidad como un modelo base de visión, pero su capacidad para controlar la granularidad de la segmentación es limitada, lo que a menudo requiere que los usuarios realicen ajustes manuales. Para abordar este desafío, se ha introducido UnSAMv2, que permite segmentar cualquier cosa a cualquier granularidad sin anotaciones humanas. Este modelo amplía la estrategia de dividir y conquistar de su predecesor, UnSAM, al descubrir numerosas parejas de máscaras y granularidades e introducir un nuevo embebido de control de granularidad para un manejo preciso …

Le modèle Segment Anything (SAM) est devenu populaire en tant que modèle de base en vision, mais il a des difficultés à contrôler la granularité de la segmentation, nécessitant souvent un raffinement manuel par les utilisateurs. Pour surmonter ce défi, UnSAMv2 a été introduit, permettant la segmentation à n'importe quelle granularité sans annotations humaines. Ce modèle s'appuie sur la stratégie de division et conquête de son prédécesseur, UnSAM, en identifiant de nombreuses paires de masques et de granularités et en mettant en œuvre un nouvel encodage de contrôle de granularité pour une gesti…

The Segment Anything Model (SAM) has gained popularity as a vision foundation model, but it struggles with controlling segmentation granularity, often requiring manual refinement by users. To overcome this challenge, UnSAMv2 has been introduced, allowing segmentation at any granularity without human annotations. This model builds on the divide-and-conquer strategy of its predecessor, UnSAM, by identifying numerous mask-granularity pairs and implementing a new granularity control embedding for precise segmentation scale management. The model demonstrates effectiveness with only 6,000 unlabeled …

UnSAMv2: Self-Supervised Learning Enables Segment Anything at Any Granularity

arXiv:2601.07970v1 Announce Type: new 
Abstract: This paper presents the Sesame Plant Segmentation Dataset, an open source annotated image dataset designed to support the development of artificial intelligence models for agricultural applications, with a specific focus on sesame plants. The dataset comprises 206 training images, 43 validation images, and 43 test images in YOLO compatible segmentation format, capturing sesame plants at early growth stages under varying environmental conditions. Data were collected using a high resolution mobile camera from farms in Jirdede, Daura Local Government Area, Katsina State, Nigeria, and annotated using the Segment Anything Model version 2 with farmer supervision. Unlike conventional bounding box datasets, this dataset employs pixel level segmentation to enable more precise detection and analysis of sesame plants in real world farm settings. Model evaluation using the Ultralytics YOLOv8 framework demonstrated strong performance for both detection and segmentation tasks. For bounding box detection, the model achieved a recall of 79 percent, precision of 79 percent, mean average precision at IoU 0.50 of 84 percent, and mean average precision from 0.50 to 0.95 of 58 percent. For segmentation, it achieved a recall of 82 percent, precision of 77 percent, mean average precision at IoU 0.50 of 84 percent, and mean average precision from 0.50 to 0.95 of 52 percent. The dataset represents a novel contribution to sesame focused agricultural vision datasets in Nigeria and supports applications such as plant monitoring, yield estimation, and agricultural research.

تم تقديم مجموعة بيانات جديدة، وهي مجموعة بيانات تقسيم نبات السمسم، التي تتضمن 206 صورة تدريب، و43 صورة تحقق، و43 صورة اختبار بتنسيق تقسيم متوافق مع YOLO. تركز هذه المجموعة على نباتات السمسم في مراحل النمو المبكرة، التي تم التقاطها تحت ظروف بيئية متنوعة في نيجيريا، وتمت معالجتها باستخدام نموذج Segment Anything الإصدار 2.

Se ha presentado un nuevo conjunto de datos, el Sesame Plant Segmentation Dataset, que incluye 206 imágenes de entrenamiento, 43 imágenes de validación y 43 imágenes de prueba formateadas para la segmentación YOLO. Este conjunto de datos se centra en las plantas de sésamo en etapas tempranas de crecimiento, capturadas bajo diversas condiciones ambientales en Nigeria, y anotadas con el Segment Anything Model versión 2.

Un nouveau jeu de données, le Sesame Plant Segmentation Dataset, a été introduit, comprenant 206 images d'entraînement, 43 images de validation et 43 images de test formatées pour la segmentation YOLO. Ce jeu de données se concentre sur les plantes de sésame à des stades précoces de croissance, capturées sous diverses conditions environnementales au Nigeria, et annotées avec le Segment Anything Model version 2.

A new dataset, the Sesame Plant Segmentation Dataset, has been introduced, featuring 206 training images, 43 validation images, and 43 test images formatted for YOLO segmentation. This dataset focuses on sesame plants at early growth stages, captured under various environmental conditions in Nigeria, and annotated with the Segment Anything Model version 2.

Sesame Plant Segmentation Dataset: A YOLO Formatted Annotated Dataset

One More Thing in AI – Your Shortcut to AI Mastery

UnSAMv2: Self-Supervised Learning Enables Segment Anything at Any Granularity

Was this article worth reading? Share it

One More Thing in AI

LucidQuery AI

Supametas.AI

VideoDigest

Zemith-3bda3b

The Visualizer

Ready to build your own newsroom?