Skip to content
Preprint

Beyond Visual Grasping: Benchmarking Complex Grasping from Detection to Execution

Jul 2026 · 0 citations · 51 references
Computer Science

TL;DR

GCA-Bench is proposed, a benchmark featuring challenging grasping with complex action scenarios that involve both scene-level reasoning and semantic constraints and enables the evaluation of recent large foundation models under the same settings.

Abstract

Robust robotic grasping remains a fundamental challenge for complex real-world applications. Recent advances in large-scale models demonstrate promising capabilities for reasoning in robotic tasks. However, existing benchmarks for grasping primarily focus on isolated, visual-based grasp pose detection, failing to capture the complexity of grasping tasks that require multi-step reasoning and semantic understanding during execution. To address this gap, we propose GCA-Bench, a benchmark featuring challenging \textit{grasping with complex action} scenarios that involve both scene-level reasoning and semantic constraints. GCA-Bench enables the evaluation of recent large foundation models under the same settings. To demonstrate the effectiveness of our new benchmark, we implement a diverse set of baselines, ranging from traditional grasp detection pipelines to end-to-end learning methods. Empirical studies achieve success rates below 70\% on complex grasping scenarios, underscoring critical limitations. In addition, we propose new evaluation metrics, analyze critical failure models, and provide insights to guide the development of more robust and generalizable grasping strategies.

View source

Similar papers

Preprint Aug 2026

GOAG: Generative and Object-Agnostic Grasp Planner for Dexterous Robotic Manipulation

Multifingered grasping is a crucial robotic skill, but current deep-learning grasp planners often struggle to generalize to new objects because they are trained on limited, object-specific datasets. We introduce a fundamentally different approach, grounded in the observation that the gripper and the object share identical surface geometry at their mutual contact points. We propose GOAG: Generative and Object-Agnostic Grasp Planner for Dexterous Robotic Manipulation, a novel deep generative model that learns a compact latent representation of a specific gripper's contact surface distribution, enabling the efficient sampling of valid grasp configurations without relying on object-specific training data. We show that by introducing object features only at inference time, our model can effectively retrieve admissible contact areas that are compatible with the gripper's capabilities. We validate our approach through extensive experiments on established grasp protocols in both simulated and real-world scenarios, demonstrating its effectiveness with different grippers from the literature. Our method delivers state-of-the-art results on the objects from the MultiDex dataset, achieving an average success rate of 86.93%. It offers significantly faster processing when generating numerous grasps, while matching the performance of leading approaches specifically trained on this dataset. Unlike these methods, our approach does not rely on object-specific training data, highlighting the advantages of object-agnostic learning. It effectively addresses the generalization challenges faced by traditional data-driven grasp planners. Code and videos are available on our project website https://cea-list.github.io/goagweb/ .

Julien Mérand, Boris Meden, Mathieu Grossard et al. · 1 citation
Preprint Jul 2026

SeededGrasp: Language-Guided Grasping in Complex Scenes with Multiple Embodiments

Practical robotic grasping in complex scenes requires both 3D spatial reasoning and alignment with task-specific requirements. Vision-language models (VLMs) offer a natural way to specify these requirements using language, but existing approaches either use a VLM to predict the grasp directly with limited spatial awareness, or train the VLM together with the grasping model, which requires significantly more data and compute. These limitations impede performance and have prevented scaling to multiple embodiments in complex scenes. We address this by proposing SeededGrasp, a novel data-efficient framework that enables a VLM to predict a seed point to be used as conditioning for a subsequent lightweight grasp-generation model. Our architecture decouples high-level semantic reasoning from low-level geometric execution, enabling multi-embodiment support while bypassing the need for expensive end-to-end training. To enable training such models, we release the first multi-embodiment tabletop grasping dataset comprising over 2.5M grasps in cluttered scenes. Experimental results demonstrate that our approach outperforms existing baselines, achieving 72% success in simulation and 78% in real-world grasping experiments. See our project site for data and code: https://uoft-isl.github.io/seeded-grasp/

Yang Xu, Gurpreet Singh Mukker, Raymond Wang et al. · 0 citations
#artificial intelligence Preprint Aug 2026

Iterative Grasp Pose Refinement: A Deep Reinforcement Learning Approach for 2D Vision

A reinforcement learning-based framework for robotic grasp refinement, integrating keypoint-based object representations with a Deep Q-Network (DQN), is proposed, offering a scalable and adaptable solution for contact-rich manipulation tasks.

Amir Arsalan Nematollahi, Shayan Ahmadi, M. T. Masouleh et al. · 0 citations
Preprint Jul 2026

IMBench: A Benchmark for Intuitive Robotic Manipulation

Humans combine reasoning and motor control to solve complex manipulation tasks under diverse constraints. They build an understanding of the physical world that helps them convert reasoning into actions and quickly adapt to new scenes, tasks, and rules. We refer to this capability as intuitive manipulation. Existing benchmarks fail to capture this integration: they evaluate physical reasoning in isolation from execution, or measure policy performance without requiring explicit reasoning. We introduce IMBENCH, a benchmark designed to evaluate intuitive manipulation as an integrated capability spanning perception, physical reasoning, action generation, and iterative execution. Our tasks require models to infer task-relevant physical structure and generate feasible action sequences under explicit constraints, including contact-rich manipulation, tool use, and multi-stage dependencies. We introduce a benchmark of 35 tasks, 14K filtered trajectories, and scalable tools for generating diverse scenarios. Experiments reveal a consistent gap: vision language models show partial physical reasoning ability but fail to produce executable plans, while state-of-the-art vision-language-action models struggle to satisfy task constraints and generalize across scenarios. These results identify intuitive manipulation as a missing axis in current foundation models and generalist robot policies, and position IMBENCH as a step toward evaluating and enabling more integrated, adaptive physical intelligence.

Anurag Maurya, Sukhvansh Jain, Prajwal Avhad et al. · 0 citations
Preprint Aug 2026

Real-World Cooperative Bimanual Dexterous Grasp of Large Objects from Single-View Observations

This work proposes a real-world bimanual grasping framework that includes a multimodal dataset capturing joint angles, visual observations and force signals; a Denoising Diffusion Probabilistic Model (DDPM)-based module that generates joint-level grasp configurations from segmented point clouds; and an execution strategy that integrates motion planning with online grasp refinement to ensure physical stability and feasibility.

Ziming Li, Mingxuan Wu, Jiaqi Zhang et al. · 0 citations
Conference Jul 2026

Evaluation of Vision-Language Models for Task-Oriented Robotic Grasping

This study comparatively examines the task-oriented grasping problem, which is of critical importance in robotic manipulation, through modern Vision-Language Models. Within the scope of the study, the performance rates of the GraspMolmo model, specifically trained for robotic tasks, and Gemini ER-1.5, a general-purpose multimodal AI, were analyzed. The evaluation process was conducted using the TaskGrasp-Image dataset, which encompasses a wide range of objects and tasks, through natural language commands and RGB-D images. The accuracy of the grasping coordinates generated by the models was systematically assessed across varying tolerance thresholds, revealing that both models achieved high accuracy rates. The robotics-specialized model demonstrated a notable advantage over the general-purpose model, particularly under strict tolerance conditions. Error analysis showed that the majority of failed predictions targeted functionally incorrect regions of the object rather than falling outside it entirely, indicating that semantic reasoning rather than geometric localization constitutes the primary challenge.

Ceren Dinç, Ayhan Küçükmanísa, Ozan Berk Kaya · 0 citations