Home > News> company news


Gift toys speech recognition applications - Xinditai

26 Oct 2017 -

First, the concept of speech recognitionOn the understanding of the speech recognition is the technology, robots can identify people speak it is an involves extensive and cross science, in the 21st century has more than 60 years of development history, speech recognition technology applied in the electronic gifts toys, let the product is the function of the richer sense of science and technology, the product can interact with people in the most natural way, not only increase the fun, more intelligent, the general speech recognition can be divided into two categories: one is the embedded speech recognition technology, the second is the cloud recognition technology.


Second, the embedded speech recognition technology and the characteristics of the cloud recognition technologyEmbedded speech recognition technology has the following major features: one is offline standalone identification;Number two is to recognize the voice entry restricted, generally between dozens to thousands of;Three is noisy environment recognition rate above 90%, algorithm, the number of entries, the specimen on the influential;Four is the noise resistance difference is big, the hardware and algorithm on its influence;Five is mostly through audio is to recognize, there are a few is to identify the text, but the effect is not ideal, is hard to see on market at present.In contrast, the cloud recognition technology has some advantages.One is the online background processing;The second is the cloud of speech recognition at present, the concept of entry, the entry does not limit in theory, and to constantly create new words;Three is noisy environment recognition rate above 95%, and with the expansion of the database, the improvement of algorithm will use more and more good;Four is environmental noise filtering, noise resistance;Five is the semantic analysis judgment ability.


Above you can see the cloud recognition technology conforms to now the application of the toy, but the cloud recognition technology and rely too much on the network, if there is no network can't use, and in operation for adults to assist children to play, of course, such a toy for now cost is quite high, is a high-end smart toys.And the embedded speech recognition technology is one of the more mature now, in the market application of less because can recognize voice entry restrictions, compared with the cloud speech recognition technology, difficult to meet the present situation of the market.New di tai electronics has developed a kind of applied in speech recognition custom scheme of electronic voice gifts toys, customize different solutions according to the different needs of customers, to meet customers and can meet the demand of the market.