“When fully developed, these architectures are expected to allow Avant! to recognize objects without the need for human intervention.” Danny Rittman, CTO, Gopher Protocol

San Diego-based Gopher Protocol has announced its entry into the field of AI-driven object recognition, with the company starting work on a system called “Avant!”. The system will be based on a Deep Neural Network, and will be trained to classify and identify objects based primarily on geometry.

In a statement announcing the project, Gopher Protocol Chief Technology Officer Danny Rittman said the Avant! engine will use a “Deformable Part-based system” that will feature a “broad class of proprietary detection algorithms”, which will enable the system to build “high-precision part-based models” for a range of object classes. “When fully developed, these architectures are expected to allow Avant! to recognize objects without the need for human intervention,” Rittman said.

Gopher Protocol’s project offers another example of the growing excitement over artificial intelligence technology, particularly with the rise of machine learning systems that can refine their algorithms on troves of data. In announcing its work on Avant!, Gopher Protocol did not offer a specific game plan for its technology in terms of applications, but noted that “having more precise and detailed object recognition is a crucial task especially for autonomous driving, anatomical/facial recognition and robotics.”