发明名称 | Sequence transcription with deep neural networks | ||
摘要 | Systems and methods for sequence transcription with neural networks are provided. More particularly, a neural network can be implemented to map a plurality of training images received by the neural network into a probabilistic model of sequences comprising P(S|X) by maximizing log P(S|X) on the plurality of training images. X represents an input image and S represents an output sequence of characters for the input image. The trained neural network can process a received image containing characters associated with building numbers. The trained neural network can generate a predicted sequence of characters by processing the received image. | ||
申请公布号 | US9454714(B1) | 申请公布日期 | 2016.09.27 |
申请号 | US201414587088 | 申请日期 | 2014.12.31 |
申请人 | Google Inc. | 发明人 | Ibarz Julian;Bulatov Yaroslav;Goodfellow Ian |
分类号 | G06K9/62 | 主分类号 | G06K9/62 |
代理机构 | Dority & Manning, P.A. | 代理人 | Dority & Manning, P.A. |
主权项 | 1. A computer-implemented method comprising: receiving, by the one or more computing devices, an image containing characters associated with character sequences; processing, by the one or more computing devices, the received image using a trained neural network, wherein the trained neural network has been trained on a plurality of training images to predict character sequences in images by maximizing log P(S|X), wherein X represents an input image and S represents an output sequence of characters for the input image; andthe plurality of training images each contain a sequence of characters having a character sequence length that is greater than 1, wherein each character in the character sequence is a discrete variable having a finite number of multiple possible values; and generating, by the one or more computing devices, with the trained neural network a predicted sequence of characters based at least in part on the processing of the received image. | ||
地址 | Mountain View CA US |