Lightweight Language-driven Grasp Detection using Conditional Consistency Model

TL;DR

This paper introduces a lightweight, language-driven grasp detection method using a conditional consistency model with diffusion processes, enabling fast inference and improved accuracy in robotic grasping tasks.

Contribution

The paper proposes a novel lightweight diffusion-based approach that integrates language prompts with grasp detection, reducing inference time and enhancing versatility.

Findings

Outperforms recent grasp detection methods in accuracy

Achieves faster inference time compared to traditional diffusion models

Validated in real-world robotic experiments

Abstract

Language-driven grasp detection is a fundamental yet challenging task in robotics with various industrial applications. In this work, we present a new approach for language-driven grasp detection that leverages the concept of lightweight diffusion models to achieve fast inference time. By integrating diffusion processes with grasping prompts in natural language, our method can effectively encode visual and textual information, enabling more accurate and versatile grasp positioning that aligns well with the text query. To overcome the long inference time problem in diffusion models, we leverage the image and text features as the condition in the consistency model to reduce the number of denoising timesteps during inference. The intensive experimental results show that our method outperforms other recent grasp detection methods and lightweight diffusion models by a clear margin. We further validate our method in real-world robotic experiments to demonstrate its fast inference time capability.