Artificial intelligence speech recognition integrated circuit chip
The chip is designed based on ASIC architecture for automatic speech recognition application. Embedded with Chipintelli's Brain Network Processing Unit(BNPU), CI1006 canperfectly support the parallel computing of the general-structure DNN with very high efficiency and quite low power-consumption. So it is quite suitable to the local speech processing and recognition by the means of deep learning for the intelligent end-products.
CI1006 has built-in MCU kernel and the general I/O interfaces for control and communication, which can support the various solutions of many kinds of intelligent end-products. So it can be widely applied to intelligent household electrical appliances, robots, intelligent toys, and many other products. It supports the large vocabulary local ASR, and it overcomes the low response, low security and even breakdown issues due to the latency, jam, drop-line of network if only via the cloud computing.
Product advantage
1. Using the independent core patent technology of artificial neural network hardware structure, which is equivalent to the super NN parallel computing power of dozens of CPU cores.
2.By adopting independent research and development of brain neural network processing unit (BNPU) core and low power design, the power consumption is about 1/10 of the same multi-core application processor chip.
3.Compared with the mainstream cloud recognition, local speech recognition has many advantages, such as fast recognition speed, privacy protection, no dependent on network, and good sense of experience.
Dual microphone intelligent speech recognition solution
The CI1006+CX20921 chip's far field dual MIC speech recognition solution is the first dual MIC program to support 10 meters faraway speech in the industry.
The feature of the solution:
1. power consumption of about 0.4 watts, conforming to national energy efficiency standards of electronic products;
2. With powerful intelligence computing ability, from command to execution, it takes only 200-800ms, and the recognition rate is up to 97%.
3. I t has the function of echo cancellation.
Single microphone intelligent speech recognition solution
In August 2017, after strict quality testing and expert customer experience, the world's first CI1006 single chip, single MIC far-field speech recognition solution has become the first MIC in the industrywhich supports 10 meters faraway speech, and its cost performance has been further improved.
The feature of the scheme:
1. power consumption is about 0.1 watts, and it is fully adapted to the national level energy efficiency standards of electronic products.
2. strong intelligent computing, from command to execution , it needs only 200-800ms, the recognition rate is up to 95%.
application areas
Consumer electronics, communications electronics, automotive electronics, medical electronics, industrial electronics.
Products have been used in Midea, Haier, Vatti, GREE, Changhong, TCL, etc. high electronics industry field.
Sichuan Jinfengli International Trade Co., Ltd.
- Phone: +86 028-85326007
- Business Type: Trading Company
- Location: China
- Main Products: facial recongnition products and service, artifacial intelligence products and service, building material
- Total Employees:100-500 people
- Year Established:2017
- Export Markets:
- Total Annual Revenue: