IDEA #62XARP Human-Inspired Multimedia Annotation System and Method

INACTIVE, combined with T002435: 22A0026. The invention includes a system and method for a human-inspired annotation system to curate multimedia data (text, graphics, animation, images, and videos). The curation and annotation are based on the human’s perception and reasoning of the data. The system employs the use of different sensors such as eye-trackers, audio, and manual annotation as a single entity to curate data. This process of data annotation mimics the perception of a human expert and presents credible and meaningful information for a wide range of applications in Artificial Intelligence. The data acquired from this system can further be used to train AI models to mimic the human expert. Further, these models can be used as a learning tool to train new personnel. The present invention helps overcome problems such as, for example, reduction in cost and time required for analyzing the multi-modal data, increasing the efficiency and accuracy of the existing and newly developed methods, and simplified usage requirements for end users. The annotation system has been developed and used to create a dataset for dental panoramic images. The findings from this work will be submitted for publication by the end of March.
For more information or to license this innovation: