Deep learning

Deep learning is a very popular technique in AI. Our experts are experienced in using Convolutional Nerual Networks, Attention-based Neural Networks, Transfer learning techniques, Generative Adversarial Networks and LSTM such as GRU for time series signal processing etc. Applications are mostly developed on Nvidia TensorRT, Tensorflow and Pytorch.
The following are some of the applications completed:

Depth Imaging

More complex computer vision applications require highly accurate 3D depth information which will be very costly using high-end structured light camera. We provide cost effective solution using our AI algorithm to compute accurate depth map using only stereo images.

Tracking

Tracking objects are very common applications. To track objects robustly and accurately is a challenge. AI enable us to track objects with state-of-the-art performance.

Object Recognition

AI allows objects to be recognized easily. In addition, it can classify different varieties of object class. For example, cars can be recognized together with its color and model.

Energy Disaggregation

Successful Grid energy deployment requires big data analytics of energy consumptions. We are able to efficiently disaggregate energy consumptions into its components using state-of-the-art time series LSTM algorithm.