408
DENSO and Toshiba will develop artificial intelligence for vehicles
Official press release two well-known Japanese company — manufacturer of spare parts and components DENSO Corporation and developed by Toshiba electronics Corporation, has announced an agreement to jointly develop artificial intelligence for cars. The joint project is called "Deep Neural Network-Intellectual Property" (DNN-IP). The final development will include self-developed by both companies technologies for object recognition that will help to create a driver assistance system and lead to the emergence of a technology of automatic driving.
As the name of the project, technology of object recognition systems in cars is based on the simulation work of the human brain. This deep learning algorithm based on neural network. In the future, the system based on Deep Neural Network needs to work better — faster and more accurately than capable of it.
The concept of the system for automatic object recognition (Toshiba)
A modern system of learning machines are based on the analysis provided by the camera of the vehicle images based on previously uploaded images. Obviously, in this case to consider all possible options for the traffic situation is simply impossible. System deep learning on the basis of neural networks able to learn from data generated by the system in the scanning space. Thus, the list of identifiable objects is expanding rapidly, and the recognition accuracy is increasing.
The partner companies DENSO and Toshiba plan to create a whole range of solutions for deep machine learning using neural networks. The block to learning can be so small that it can be embedded in the part of the processor for vehicle control or car surveillance camera.
The control units systems for recognition of training will be produced by the company DENSO (DENSO)
A comprehensive system to assist the driver in steering the car or the autopilot will be developed by the company DENSO. Toshiba takes on the challenge to implement the technology of "smart" recognition of objects on the road in the form of electronic circuits. The effectiveness of specialized solutions DNN-IP promises to be more widely used in the field of the universal digital signal processors or graphics cards. published
Source: www.3dnews.ru/941158
As the name of the project, technology of object recognition systems in cars is based on the simulation work of the human brain. This deep learning algorithm based on neural network. In the future, the system based on Deep Neural Network needs to work better — faster and more accurately than capable of it.
The concept of the system for automatic object recognition (Toshiba)
A modern system of learning machines are based on the analysis provided by the camera of the vehicle images based on previously uploaded images. Obviously, in this case to consider all possible options for the traffic situation is simply impossible. System deep learning on the basis of neural networks able to learn from data generated by the system in the scanning space. Thus, the list of identifiable objects is expanding rapidly, and the recognition accuracy is increasing.
The partner companies DENSO and Toshiba plan to create a whole range of solutions for deep machine learning using neural networks. The block to learning can be so small that it can be embedded in the part of the processor for vehicle control or car surveillance camera.
The control units systems for recognition of training will be produced by the company DENSO (DENSO)
A comprehensive system to assist the driver in steering the car or the autopilot will be developed by the company DENSO. Toshiba takes on the challenge to implement the technology of "smart" recognition of objects on the road in the form of electronic circuits. The effectiveness of specialized solutions DNN-IP promises to be more widely used in the field of the universal digital signal processors or graphics cards. published
Source: www.3dnews.ru/941158
