Technologies used by VideoCurio
Technologies used by VideoCurio include computer vision with Deep Neural Network for object detection and classification. VideoCurio3 also use the power of Nvidia GPUs and their Cuda cores for fast video encoding and decoding.
NVIDIA GPU
VideoCurio 3.0 rely on NVIDIA CUDA technology on GPU for fast video encoding and decoding. It also benefit from TensorRT cores for accelerating VideoCurio3 neural networks.
TensorFlow
VideoCurio 3.0 neural networks are tested, trained, developed and run under TensorFlow.
Google cloud platform
VideoCurio Apps are hosted on the Google cloud platform and use some of it's products like App Engine and Google Cloud bucket storage. Some tests are made with the Google AI platform and Tensorflow Lite.
Open Source softwares
At VideoCurio we are proud users of Open Source softwares, including but not limited to OpenCV, FFmpeg, Python and Ubuntu. FFmpeg and OpenCV are compiled by us from sources to include NVIDIA Deep Neural Network support and NVIDIA video codec SDK.