arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1702.07975cs.CV
收藏

构造快速紧凑的卷积神经网络用于脱机手写体汉字识别

Building Fast and Compact Convolutional Neural Networks for Offline Handwritten Chinese Character Recognition

South China University of Technology(华南理工大学);Fujitsu Research & Development Center Co. Ltd.(富士通研究开发中心有限公司)

Xuefeng Xiao, Lianwen Jin, Yafeng Yang, Weixin Yang, Jun Sun, Tianhai Chang

摘要:

与计算机视觉中的其他问题一样,离线手写汉字识别(HCCR)使用基于卷积神经网络(CNN)的方法取得了令人印象深刻的结果。然而,需要更大和更深的网络来提供该领域的最新成果。这样的网络直观地表现为产生高计算成本,并且需要存储大量参数,这使得它们不可行地部署在便携式设备中。为了解决这个问题,我们提出了一种全局监督低秩扩展(GSLRE)方法和自适应降权(ADW)技术来解决速度和存储容量的问题。我们为HCCR设计了一个由3,755个类组成的九层CNN,并设计了一种算法,可以将网络计算成本降低9倍,并将网络压缩到基线模型原始大小的1/18,准确度仅下降0.21%。在测试中,该算法超过了文献中迄今为止报道的最佳单网络性能,同时只需要2.3 MB的存储空间。此外,当与我们有效的前向实现集成时,离线字符图像的识别在CPU上仅需9.7 ms。与HCCR的最先进的CNN模型相比,我们的方法大约快30倍,但成本效益高10倍。

英文摘要:

Like other problems in computer vision, offline handwritten Chinese character recognition (HCCR) has achieved impressive results using convolutional neural network (CNN)-based methods. However, larger and deeper networks are needed to deliver state-of-the-art results in this domain. Such networks intuitively appear to incur high computational cost, and require the storage of a large number of parameters, which renders them unfeasible for deployment in portable devices. To solve this problem, we propose a Global Supervised Low-rank Expansion (GSLRE) method and an Adaptive Drop-weight (ADW) technique to solve the problems of speed and storage capacity. We design a nine-layer CNN for HCCR consisting of 3,755 classes, and devise an algorithm that can reduce the networks computational cost by nine times and compress the network to 1/18 of the original size of the baseline model, with only a 0.21% drop in accuracy. In tests, the proposed algorithm surpassed the best single-network performance reported thus far in the literature while requiring only 2.3 MB for storage. Furthermore, when integrated with our effective forward implementation, the recognition of an offline character image took only 9.7 ms on a CPU. Compared with the state-of-the-art CNN model for HCCR, our approach is approximately 30 times faster, yet 10 times more cost efficient.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑