arXiv:2510.24563v2 Announce Type: replace 
Abstract: With advances in decision-making and reasoning capabilities, multimodal agents show strong potential in computer application scenarios. Past evaluations have mainly assessed GUI interaction skills, while tool invocation abilities, such as those enabled by the Model Context Protocol (MCP), have been largely overlooked. Comparing agents with integrated tool invocation to those evaluated only on GUI interaction is inherently unfair. We present OSWorld-MCP, the first comprehensive and fair benchmark for assessing computer-use agents' tool invocation, GUI operation, and decision-making abilities in a real-world environment. We design a novel automated code-generation pipeline to create tools and combine them with a curated selection from existing tools. Rigorous manual validation yields 158 high-quality tools (covering 7 common applications), each verified for correct functionality, practical applicability, and versatility. Extensive evaluations of state-of-the-art multimodal agents on OSWorld-MCP show that MCP tools generally improve task success rates (e.g., from 8.3% to 20.4% for OpenAI o3 at 15 steps, from 40.1% to 43.3% for Claude 4 Sonnet at 50 steps), underscoring the importance of assessing tool invocation capabilities. However, even the strongest models have relatively low tool invocation rates, Only 36.3%, indicating room for improvement and highlighting the benchmark's challenge. By explicitly measuring MCP tool usage skills, OSWorld-MCP deepens understanding of multimodal agents and sets a new standard for evaluating performance in complex, tool-assisted environments. Our code, environment, and data are publicly available at https://osworld-mcp.github.io.

تم تقديم معيار OSWorld-MCP لتقييم قدرات استدعاء الأدوات لوكلاء استخدام الكمبيوتر بشكل عادل إلى جانب عمليات واجهة المستخدم الرسومية. يتضمن هذا المعيار 158 أداة تم التحقق منها عبر سبع تطبيقات، مما يظهر تحسينات كبيرة في معدلات نجاح المهام لوكلاء مثل OpenAI o3 وClaude 4 Sonnet. إن تقديم هذا المعيار أمر بالغ الأهمية لأنه يعالج الجانب الذي تم تجاهله سابقًا من استدعاء الأدوات في الوكلاء متعدد الوسائط، مما يبرز إمكاناتهم في التطبيقات الواقعية.

Se introdujo el benchmark OSWorld-MCP para evaluar de manera justa las capacidades de invocación de herramientas de los agentes de uso informático junto con las operaciones de GUI. Este benchmark incluye 158 herramientas validadas en siete aplicaciones, mostrando mejoras significativas en las tasas de éxito de tareas para agentes como OpenAI o3 y Claude 4 Sonnet. La introducción de este benchmark es crucial, ya que aborda el aspecto previamente pasado por alto de la invocación de herramientas en agentes multimodales, destacando su potencial en aplicaciones del mundo real.

Le benchmark OSWorld-MCP a été introduit pour évaluer équitablement les capacités d'invocation d'outils des agents d'utilisation informatique aux côtés des opérations GUI. Ce benchmark comprend 158 outils validés couvrant sept applications, montrant des améliorations significatives des taux de réussite des tâches pour des agents comme OpenAI o3 et Claude 4 Sonnet. L'introduction de ce benchmark est cruciale car elle aborde l'aspect négligé de l'invocation d'outils chez les agents multimodaux, soulignant leur potentiel dans des applications réelles.

The OSWorld-MCP benchmark was introduced to fairly evaluate computer-use agents' tool invocation capabilities alongside GUI operations. This benchmark includes 158 validated tools across seven applications, demonstrating significant improvements in task success rates for agents like OpenAI o3 and Claude 4 Sonnet. The introduction of this benchmark is crucial as it addresses the previously overlooked aspect of tool invocation in multimodal agents, highlighting their potential in real-world applications.

OSWorld-MCP: Benchmarking MCP Tool Invocation In Computer-Use Agents

arXiv:2511.14109v1 Announce Type: new 
Abstract: Visual Place Recognition (VPR) aims to match query images against a database using visual cues. State-of-the-art methods aggregate features from deep backbones to form global descriptors. Optimal transport-based aggregation methods reformulate feature-to-cluster assignment as a transport problem, but the standard Sinkhorn algorithm symmetrically treats source and target marginals, limiting effectiveness when image features and cluster centers exhibit substantially different distributions. We propose an asymmetric aggregation VPR method with geometric constraints for locally aggregated descriptors, called $A^2$GC-VPR. Our method employs row-column normalization averaging with separate marginal calibration, enabling asymmetric matching that adapts to distributional discrepancies in visual place recognition. Geometric constraints are incorporated through learnable coordinate embeddings, computing compatibility scores fused with feature similarities, thereby promoting spatially proximal features to the same cluster and enhancing spatial awareness. Experimental results on MSLS, NordLand, and Pittsburgh datasets demonstrate superior performance, validating the effectiveness of our approach in improving matching accuracy and robustness.

$A^2$GC-VPR هو أسلوب جديد للتعرف على الأماكن البصرية (VPR) يتناول قيود أساليب التجميع التقليدية في مطابقة صور الاستعلام مع قاعدة بيانات. من خلال اعتماد نهج تجميع غير متماثل مع قيود هندسية، يعزز هذا الأسلوب فعالية مطابقة الميزات، خاصة عند التعامل مع توزيعات متباينة لميزات الصورة ومراكز التجمع. تستخدم التقنية متوسطات تطبيع الصفوف والأعمدة مع تضمينات إحداثيات قابلة للتعلم لتحسين درجات التوافق لوصفيات التجميع المحلي.

$A^2$GC-VPR es un nuevo método para el Reconocimiento Visual de Lugares (VPR) que aborda las limitaciones de los métodos de agregación tradicionales al emparejar imágenes de consulta con una base de datos. Al emplear un enfoque de agregación asimétrica con restricciones geométricas, este método mejora la efectividad del emparejamiento de características, especialmente cuando se enfrentan a distribuciones variables de características de imagen y centros de clúster. La técnica utiliza promedios de normalización fila-columna y embeddings de coordenadas aprendibles para mejorar las puntuaciones de…

$A^2$GC-VPR est une nouvelle méthode pour la reconnaissance de lieux visuels (VPR) qui s'attaque aux limites des méthodes d'agrégation traditionnelles dans l'appariement d'images de requête à une base de données. En adoptant une approche d'agrégation asymétrique avec des contraintes géométriques, cette méthode améliore l'efficacité de l'appariement des caractéristiques, en particulier lorsqu'il s'agit de distributions variées des caractéristiques d'image et des centres de clusters. La technique utilise une moyenne de normalisation ligne-colonne et des embeddings de coordonnées apprenables pour…

$A^2$GC-VPR is a new method for Visual Place Recognition (VPR) that addresses the limitations of traditional aggregation methods in matching query images to a database. By employing an asymmetric aggregation approach with geometric constraints, this method enhances the effectiveness of feature matching, particularly when dealing with varying distributions of image features and cluster centers. The technique utilizes row-column normalization averaging and learnable coordinate embeddings to improve compatibility scores for locally aggregated descriptors.

$A^2$GC: $A$symmetric $A$ggregation with Geometric Constraints for Locally Aggregated Descriptors

arXiv:2511.13911v1 Announce Type: new 
Abstract: Despite recent progress in predicting biomarker trajectories from real clinical data, uncertainty in the predictions poses high-stakes risks (e.g., misdiagnosis) that limit their clinical deployment. To enable safe and reliable use of such predictions in healthcare, we introduce a conformal method for uncertainty-calibrated prediction of biomarker trajectories resulting from randomly-timed clinical visits of patients. Our approach extends conformal prediction to the setting of randomly-timed trajectories via a novel nonconformity score that produces prediction bands guaranteed to cover the unknown biomarker trajectories with a user-prescribed probability. We apply our method across a wide range of standard and state-of-the-art predictors for two well-established brain biomarkers of Alzheimer's disease, using neuroimaging data from real clinical studies. We observe that our conformal prediction bands consistently achieve the desired coverage, while also being tighter than baseline prediction bands. To further account for population heterogeneity, we develop group-conditional conformal bands and test their coverage guarantees across various demographic and clinically relevant subpopulations. Moreover, we demonstrate the clinical utility of our conformal bands in identifying subjects at high risk of progression to Alzheimer's disease. Specifically, we introduce an uncertainty-calibrated risk score that enables the identification of 17.5% more high-risk subjects compared to standard risk scores, highlighting the value of uncertainty calibration in real-world clinical decision making. Our code is available at github.com/vatass/ConformalBiomarkerTrajectories.

تم تطوير طريقة جديدة للتنبؤ بمسارات المؤشرات الحيوية في مرض الزهايمر، حيث تعالج عدم اليقين في التنبؤات السريرية. تتيح هذه التقنية التنبؤية المتوافقة إجراء تنبؤات موثوقة من الزيارات السريرية التي تتم في أوقات عشوائية، مما يضمن أن تغطي نطاقات التنبؤ المسارات الفعلية للمؤشرات الحيوية باحتمالية محددة. تم اختبار النهج باستخدام بيانات التصوير العصبي من دراسات سريرية معروفة، مما يوضح فعاليته في تعزيز أمان التطبيقات السريرية.

Se ha desarrollado un nuevo método para predecir las trayectorias de biomarcadores en la enfermedad de Alzheimer, abordando la incertidumbre en las predicciones clínicas. Esta técnica de predicción conforme permite pronósticos fiables a partir de visitas clínicas programadas aleatoriamente, asegurando que las bandas de predicción cubran las trayectorias reales de los biomarcadores con una probabilidad especificada. El enfoque se probó utilizando datos de neuroimagen de estudios clínicos establecidos, demostrando su eficacia para mejorar la seguridad de las aplicaciones clínicas.

Une nouvelle méthode de prédiction des trajectoires de biomarqueurs dans la maladie d'Alzheimer a été développée, abordant l'incertitude des prédictions cliniques. Cette technique de prédiction conforme permet des prévisions fiables à partir de visites cliniques chronométrées de manière aléatoire, garantissant que les bandes de prédiction couvrent les trajectoires réelles des biomarqueurs avec une probabilité spécifiée. L'approche a été testée à l'aide de données d'imagerie cérébrale provenant d'études cliniques établies, démontrant son efficacité pour améliorer la sécurité des applications cl…

A new method for predicting biomarker trajectories in Alzheimer's disease has been developed, addressing the uncertainty in clinical predictions. This conformal prediction technique allows for reliable forecasts from randomly-timed clinical visits, ensuring that the prediction bands cover the actual biomarker trajectories with a specified probability. The approach was tested using neuroimaging data from established clinical studies, demonstrating its effectiveness in enhancing the safety of clinical applications.

Uncertainty-Calibrated Prediction of Randomly-Timed Biomarker Trajectories with Conformal Bands

arXiv:2511.14021v1 Announce Type: new 
Abstract: Humans can easily identify anatomical planes (axial, coronal, and sagittal) on a 2D MRI slice, but automated systems struggle with this task. Missing plane orientation metadata can complicate analysis, increase domain shift when merging heterogeneous datasets, and reduce accuracy of diagnostic classifiers. This study develops a classifier that accurately generates plane orientation metadata. We adopt a 2.5D context-aware model that leverages multi-slice information to avoid ambiguity from isolated slices and enable robust feature learning. We train the 2.5D model on both 3D slice sequences and static 2D images. While our 2D reference model achieves 98.74% accuracy, our 2.5D method raises this to 99.49%, reducing errors by 60%, highlighting the importance of 2.5D context. We validate the utility of our generated metadata in a brain tumor detection task. A gated strategy selectively uses metadata-enhanced predictions based on uncertainty scores, boosting accuracy from 97.0% with an image-only model to 98.0%, reducing misdiagnoses by 33.3%. We integrate our plane orientation model into an interactive web application and provide it open-source.

تقدم دراسة جديدة نموذجًا ثلاثي الأبعاد 2.5 يعتمد على السياق للكشف عن اتجاهات الطائرات في التصوير بالرنين المغناطيسي، مما يعالج التحديات التي تواجهها الأنظمة الآلية في تحديد الطائرات التشريحية. يستخدم النموذج معلومات متعددة الشرائح لتحسين تعلم الميزات وتقليل الغموض، محققًا معدل دقة يبلغ 99.49% مقارنة بـ 98.74% للنماذج ثنائية الأبعاد التقليدية. تعتبر هذه الخطوة مهمة بشكل خاص لتحسين مصنّفات التشخيص وتسهيل الكشف عن الأورام الدماغية من خلال توليد بيانات وصفية دقيقة حول اتجاه الطائرات.

Un nuevo estudio presenta un modelo 2.5D consciente del contexto para la detección de orientaciones de planos en resonancias magnéticas (IRM), abordando los desafíos que enfrentan los sistemas automatizados para identificar planos anatómicos. El modelo utiliza información de múltiples cortes para mejorar el aprendizaje de características y reducir la ambigüedad, logrando una tasa de precisión del 99.49% en comparación con el 98.74% de los modelos 2D tradicionales. Este avance es particularmente significativo para mejorar los clasificadores diagnósticos y facilitar la detección de tumores cereb…

Une nouvelle étude présente un modèle 2.5D contextuel pour la détection des orientations des plans IRM, répondant aux défis rencontrés par les systèmes automatisés pour identifier les plans anatomiques. Le modèle utilise des informations multi-coupes pour améliorer l'apprentissage des caractéristiques et réduire l'ambiguïté, atteignant un taux de précision de 99,49 % contre 98,74 % pour les modèles 2D traditionnels. Cette avancée est particulièrement significative pour améliorer les classificateurs diagnostiques et faciliter la détection des tumeurs cérébrales en générant des métadonnées préci…

A new study presents a context-aware 2.5D model for detecting MRI plane orientations, addressing challenges faced by automated systems in identifying anatomical planes. The model utilizes multi-slice information to enhance feature learning and reduce ambiguity, achieving a 99.49% accuracy rate compared to 98.74% for traditional 2D models. This advancement is particularly significant for improving diagnostic classifiers and facilitating brain tumor detection by generating accurate plane orientation metadata.

MRI Plane Orientation Detection using a Context-Aware 2.5D Model

arXiv:2511.14093v1 Announce Type: new 
Abstract: Cross-view object Geo-localization aims to precisely pinpoint the same object across large-scale satellite imagery based on drone images. Due to significant differences in viewpoint and scale, coupled with complex background interference, traditional multi-stage "retrieval-matching" pipelines are prone to cumulative errors. To address this, we present SMGeo, a promptable end-to-end transformer-based model for object Geo-localization. This model supports click prompting and can output object Geo-localization in real time when prompted to allow for interactive use. The model employs a fully transformer-based architecture, utilizing a Swin-Transformer for joint feature encoding of both drone and satellite imagery and an anchor-free transformer detection head for coordinate regression. In order to better capture both inter-modal and intra-view dependencies, we introduce a grid-level sparse Mixture-of-Experts (GMoE) into the cross-view encoder, allowing it to adaptively activate specialized experts according to the content, scale and source of each grid. We also employ an anchor-free detection head for coordinate regression, directly predicting object locations via heat-map supervision in the reference images. This approach avoids scale bias and matching complexity introduced by predefined anchor boxes. On the drone-to-satellite task, SMGeo achieves leading performance in accuracy at IoU=0.25 and mIoU metrics (e.g., 87.51%, 62.50%, and 61.45% in the test set, respectively), significantly outperforming representative methods such as DetGeo (61.97%, 57.66%, and 54.05%, respectively). Ablation studies demonstrate complementary gains from shared encoding, query-guided fusion, and grid-level sparse mixture-of-experts.

SMGeo هو نموذج جديد يعتمد على المحولات مصمم لتحديد المواقع الجغرافية للأجسام عبر وجهات نظر مختلفة، مما يتيح تحديدًا دقيقًا لنفس الجسم عبر الصور الساتلية وصور الطائرات بدون طيار. تواجه الطرق التقليدية أخطاء تراكمية بسبب الاختلافات في وجهات النظر والمقياس. يتعامل SMGeo مع هذه التحديات من خلال بنية تعتمد بالكامل على المحولات، باستخدام Swin-Transformer لترميز الميزات ورأس كشف بدون مرساة لعملية الانحدار الإحداثي. يسمح النموذج بتحديد المواقع الجغرافية في الوقت الفعلي عند الطلب، مما يعزز الاستخدام التفاعلي.

SMGeo es un nuevo modelo basado en transformadores diseñado para la geo-localización de objetos en vista cruzada, permitiendo la identificación precisa del mismo objeto a través de imágenes satelitales y de drones. Los métodos tradicionales suelen enfrentar errores acumulativos debido a diferencias en el punto de vista y la escala. SMGeo aborda estos desafíos con una arquitectura completamente basada en transformadores, utilizando un Swin-Transformer para la codificación de características y una cabeza de detección sin ancla para la regresión de coordenadas. El modelo permite la geo-localizaci…

SMGeo est un nouveau modèle basé sur des transformateurs conçu pour la géolocalisation d'objets en vue croisée, permettant une identification précise du même objet à travers des images satellites et de drones. Les méthodes traditionnelles rencontrent souvent des erreurs cumulatives en raison des différences de point de vue et d'échelle. SMGeo répond à ces défis avec une architecture entièrement basée sur des transformateurs, utilisant un Swin-Transformer pour l'encodage des caractéristiques et une tête de détection sans ancre pour la régression des coordonnées. Le modèle permet une géolocalisa…

SMGeo is a new transformer-based model designed for cross-view object geo-localization, enabling precise identification of the same object across satellite and drone imagery. Traditional methods often face cumulative errors due to differences in viewpoint and scale. SMGeo addresses these challenges with a fully transformer architecture, utilizing a Swin-Transformer for feature encoding and an anchor-free detection head for coordinate regression. The model allows for real-time geo-localization through click prompting, enhancing interactive usability.

SMGeo: Cross-View Object Geo-Localization with Grid-Level Mixture-of-Experts

arXiv:2511.14086v1 Announce Type: new 
Abstract: Despite recent progress in 3D-LLMs, they remain limited in accurately grounding language to visual and spatial elements in 3D environments. This limitation stems in part from training data that focuses on language reasoning rather than spatial understanding due to scarce 3D resources, leaving inherent grounding biases unresolved. To address this, we propose 3D scene editing as a key mechanism to generate precise visual counterfactuals that mitigate these biases through fine-grained spatial manipulation, without requiring costly scene reconstruction or large-scale 3D data collection. Furthermore, to make these edits targeted and directly address the specific weaknesses of the model, we introduce DEER-3D, an error-driven framework following a structured "Decompose, Diagnostic Evaluation, Edit, and Re-train" workflow, rather than broadly or randomly augmenting data as in conventional approaches. Specifically, upon identifying a grounding failure of the 3D-LLM, our framework first diagnoses the exact predicate-level error (e.g., attribute or spatial relation). It then executes minimal, predicate-aligned 3D scene edits, such as recoloring or repositioning, to produce targeted counterfactual supervision for iterative model fine-tuning, significantly enhancing grounding accuracy. We evaluate our editing pipeline across multiple benchmarks for 3D grounding and scene understanding tasks, consistently demonstrating improvements across all evaluated datasets through iterative refinement. DEER-3D underscores the effectiveness of targeted, error-driven scene editing in bridging linguistic reasoning capabilities with spatial grounding in 3D LLMs.

أظهرت التقدمات الأخيرة في نماذج اللغة ثلاثية الأبعاد (3D-LLMs) قيودها في ربط اللغة بدقة بالعناصر البصرية والمكانية في البيئات ثلاثية الأبعاد. تنبع هذه المشكلة من بيانات التدريب التي تركز على التفكير اللغوي بدلاً من الفهم المكاني، مما يترك تحيزات الربط غير محلولة. لمعالجة ذلك، تم اقتراح نهج جديد يسمى DEER-3D، الذي يستخدم تحرير المشاهد المدفوع بالأخطاء لإنشاء مضادات بصرية دقيقة، مما يسمح بالتلاعب المكاني الدقيق دون الحاجة إلى إعادة بناء المشهد المكلفة.

Los recientes avances en modelos de lenguaje 3D (3D-LLMs) han puesto de manifiesto sus limitaciones para anclar con precisión el lenguaje a elementos visuales y espaciales en entornos 3D. Este problema se debe a que los datos de entrenamiento se centran en el razonamiento lingüístico en lugar de en la comprensión espacial, lo que deja sin resolver los sesgos de anclaje. Para abordar esto, se propone un nuevo enfoque llamado DEER-3D, que utiliza la edición de escenas impulsada por errores para generar contrafactuales visuales precisos, permitiendo una manipulación espacial detallada sin necesid…

Les récents progrès des modèles de langage 3D (3D-LLMs) ont mis en évidence leurs limites dans l'ancrage précis du langage aux éléments visuels et spatiaux dans les environnements 3D. Ce problème découle de données d'entraînement qui mettent l'accent sur le raisonnement linguistique plutôt que sur la compréhension spatiale, entraînant des biais d'ancrage non résolus. Pour y remédier, une nouvelle approche appelée DEER-3D est proposée, qui utilise l'édition de scènes guidée par les erreurs pour créer des contrefactuels visuels précis, permettant une manipulation spatiale fine sans reconstructio…

Recent advancements in 3D large language models (3D-LLMs) have highlighted their limitations in accurately grounding language to visual and spatial elements in 3D environments. This issue arises from training data that emphasizes language reasoning over spatial understanding, leading to unresolved grounding biases. To tackle this, a new approach called DEER-3D is proposed, which utilizes error-driven scene editing to create precise visual counterfactuals. This method allows for fine-grained spatial manipulation without the need for extensive scene reconstruction or large-scale data collection.

Error-Driven Scene Editing for 3D Grounding in Large Language Models

arXiv:2511.14107v1 Announce Type: new 
Abstract: Depth information is crucial for autonomous driving and intelligent robot navigation. The simplicity and flexibility of self-supervised monocular depth estimation are conducive to its role in these fields. However, most existing monocular depth estimation models consume many computing resources. Although some methods have reduced the model's size and improved computing efficiency, the performance deteriorates, seriously hindering the real-world deployment of self-supervised monocular depth estimation models in the real world. To address this problem, we proposed a real-time self-supervised monocular depth estimation method and implemented it in the real world. It is called RTS-Mono, which is a lightweight and efficient encoder-decoder architecture. The encoder is based on Lite-Encoder, and the decoder is designed with a multi-scale sparse fusion framework to minimize redundancy, ensure performance, and improve inference speed. RTS-Mono achieved state-of-the-art (SoTA) performance in high and low resolutions with extremely low parameter counts (3 M) in experiments based on the KITTI dataset. Compared with lightweight methods, RTS-Mono improved Abs Rel and Sq Rel by 5.6% and 9.8% at low resolution and improved Sq Rel and RMSE by 6.1% and 1.9% at high resolution. In real-world deployment experiments, RTS-Mono has extremely high accuracy and can perform real-time inference on Nvidia Jetson Orin at a speed of 49 FPS. Source code is available at https://github.com/ZYCheng777/RTS-Mono.

RTS-Mono هو طريقة جديدة لتقدير العمق الأحادي الذاتي الإشراف في الوقت الحقيقي، تهدف إلى تحسين القيادة الذاتية والتنقل الذكي للروبوتات. غالبًا ما تتطلب نماذج تقدير العمق الأحادي التقليدية موارد حوسبة كبيرة، مما يمكن أن يعيق تطبيقها العملي. RTS-Mono يعالج هذه المشكلة من خلال بنية خفيفة الوزن من نوع مشفر-مفكك، تستخدم مشفرًا خفيفًا وإطار دمج متناثر متعدد المقاييس لتحسين الأداء وسرعة الاستدلال، مما يجعلها مناسبة للنشر في العالم الحقيقي.

RTS-Mono es un nuevo método de estimación de profundidad monocular auto-supervisada en tiempo real, destinado a mejorar la conducción autónoma y la navegación robótica inteligente. Los modelos tradicionales de estimación de profundidad monocular suelen requerir muchos recursos informáticos, lo que puede obstaculizar su aplicación práctica. RTS-Mono aborda este problema con una arquitectura ligera de codificador-decodificador, utilizando un Lite-Encoder y un marco de fusión dispersa a múltiples escalas para optimizar el rendimiento y la velocidad de inferencia, haciéndolo adecuado para el despl…

RTS-Mono est une nouvelle méthode d'estimation de profondeur monoculaire auto-supervisée en temps réel, visant à améliorer la conduite autonome et la navigation robotique intelligente. Les modèles traditionnels d'estimation de profondeur monoculaire nécessitent souvent d'importantes ressources informatiques, ce qui peut freiner leur application pratique. RTS-Mono résout ce problème grâce à une architecture légère d'encodeur-décodeur, utilisant un Lite-Encoder et un cadre de fusion sparse multi-échelle pour optimiser les performances et la vitesse d'inférence, la rendant adaptée à un déploiemen…

RTS-Mono is a newly proposed real-time self-supervised monocular depth estimation method aimed at enhancing autonomous driving and intelligent robot navigation. Traditional monocular depth estimation models often require significant computing resources, which can hinder their practical application. RTS-Mono addresses this issue with a lightweight encoder-decoder architecture, utilizing a Lite-Encoder and a multi-scale sparse fusion framework to optimize performance and inference speed, making it suitable for real-world deployment.

OSWorld-MCP: Benchmarking MCP Tool Invocation In Computer-Use Agents

Was this article worth reading? Share it