Meet the COROB Open Call Winners: Interview with AI+PICK


AI+Pick is an AI-powered bin-picking solution designed for the flexible handling of multigeometry industrial parts in cluttered bins. It combines RGB images and points clouds acquisition with CAD-based perception and a weakly supervised deep learning pipeline to robustly detect parts under occlusions and challenging surfaces. The system estimates 6D poses, synthesizes and ranks feasible grasps, and generates robot trajectories.  

What makes it novel:

The AI+Pick solution is a significant innovation in the bin-picking market, as it combines two different approaches to create a unified product. AI+Pick combines a weakly supervised industrial instance-segmentation model with a foundation-model refinement stage and CADbased registration, thereby reducing the pose search space and improving reliability in cluttered environments. 


The reasons we applied to the COROB Open Call are our commitment to promoting automation through artificial intelligence-based solutions, particularly in object manipulation using robotic and industrial vision systems. Specifically, our mission is to increase the flexibility of industrial processes by transferring the latest academic research results into reliable industrial solutions. Thanks to the Open Call, we have developed an innovative solution that aims to revolutionise bin-picking activities by leveraging cutting-edge deep learning and 6D position estimation, enabling robots to interact with objects in a highly flexible and precise manner. Our project aligns with COROB’s focus on robotics innovation, and we see this opportunity as a way to expand our solution, strengthen our collaboration with academic and industrial partners, and contribute to the evolution of automation in challenging environments. Through AI+Pick, we aim to address key industry needs for more adaptive and efficient robotic systems. 


We envision AI+Pick making a significant impact by enabling robots to perform more complex and adaptive manipulation tasks across various industries, from logistics to manufacturing. By combining advanced artificial intelligence algorithms with 6D position estimation, our project enhances robot flexibility, enabling them to handle a wider range of objects. This will enable greater efficiency in automated workflows, reduce costs, and unlock new automation opportunities in industries that require high precision and adaptability, allowing human operators to focus on strategic, higher-value roles. Ultimately, AI+Pick will contribute to the spread of robotics, promoting productivity and innovation across all industrial sectors. 

Learn more about this solution at: https://www.it-robotics.it/en/ai-pick/ 

Similar Posts