Artificial Intelligence, Machine learning, deep learning, neural networks have revolutionized the field of image understanding and recognition. Mindtech is to be the go-to-partner for disruptive, AI-enabled visual processing solutions. Vast quantities of labelled video are required to train neural networks for visual applications; to manually capture and annotate this data has many limitations, including the time and expense involved. To date this has meant only a privileged few can develop and deploy their AI based computer visions solutions or are perhaps limited to the usage of a few, limited public datasets.
Mindtech changes that with Chameleon, the world’s first tool for synthetic data generation, including a complete virtual world simulator with scenario editor, targeted at all industries that wish to deploy their next generation AI based vision algorithms. The tool allows users to be confident that corner cases which are difficult or impossible to film in the real world have been tested, include sufficient diversity training and are fully compliant with GDPR.
This enhanced data for training helps to prevent embarrassing (or catastrophic) failures in the field. A wide range of fully accurate annotations, including pixel perfect masks, precise range data, and derivatives such as velocity are easily generated. And the use of synthetic data improves accuracy of neural networks, can actively reduce bias and vastly reduce the amount of ‘real’ data required, saving time and money.
Whether a company is working on drone detection, self-driving cars, smart security cameras or other applications requiring visual stimuli, the Chameleon simulator will provide you with the additional training data you need.