The Mindtech AI tools and services family comprises a comprehensive set of tools designed to create, manage and generate synthetic data for neural network training. These tools work with any industry standard framework such as Tensorflow, PyTorch and Caffe, which will be used to train the data to produce trained neural networks for any Inferencing platform such as Mindtech’s own AI & vision engine.
Artificial Intelligence, Machine learning, deep learning, neural networks… by whatever name you know the technology, they have revolutionised the field of image understanding and recognition. From the victory of the Alexnet neural network in the 2012 Imagenet Large Scale Visual Recognition Challenge (ILSVRC), to the 2019 introduction of Tesla’s FSD SoC into production vehicles, with 72 TOPs of AI compute performance these deep learning solutions have proved their value.
Neural networks consist of two building blocks: The network itself, consisting of a number of different interconnected layers performing various functions and the data to train this network. The two are inseparable; without both elements, the network will not function
Mindtech Chameleon is a solution to the large amounts of annotated data required to effectively train a neural network. The Mindtech Chameleon simulator creates unlimited quantities of annotated synthetic images to complement real world images. Mindtech’s Chameleon AI tools manage these large training datasets, including the import and merging of any real data available, and the visualisation and analysis of results.
Data can be created to specifically target bias in the real data set
Datasets can be stored and used without privacy concerns
Data can be perfectly annotated, including advanced annotations such as 3D-Bounding boxes, velocity vectors and distance data
Synthetic data can be generated to represent multiple different sensor types, simultaneously, RGB data, Bayer data, LIDAR data and thermal data are a few examples
The synthetic data allows you to have inference system independent data, and then apply the relevant biases on a case by case basis
Dangerous, difficult and time consuming corner test cases can be modelled with the synthetic data.
