首页> 外国专利> CHEMICAL INFORMATION IDENTIFICATION METHOD BASED ON DEEP LEARNING IMAGE IDENTIFICATION TECHNOLOGY

CHEMICAL INFORMATION IDENTIFICATION METHOD BASED ON DEEP LEARNING IMAGE IDENTIFICATION TECHNOLOGY

机译:基于深度学习图像识别技术的化学信息识别方法

摘要

A chemical information identification method based on deep learning image identification technology, the method comprising the following steps: (1) using a node target identifier to implement identification of an inputted image; (2) using a handwritten font target identifier to implement identification of the text content for the nodes identified in step (1) and thereby determine specific atoms corresponding to the nodes; (3) combining the identified plurality of atoms in pairs, and then using a chemical bond target identifier to implement identification of the chemical bonds between two atoms; (4) finding the attributes of the identified atoms in a database, calculating the relevant attributes of the structural formula, and outputting same; or storing the identified atoms and chemical bonds between the atoms as a custom king format file or drawing into a new image, and outputting same. The present method can solve the problem of identifying chemical structural or reaction formulae in hand drawings and pictures, and can be widely used in the daily work of chemists.
机译:一种基于深度学习图像识别技术的化学信息识别方法,该方法包括以下步骤:(1)使用节点目标标识符对输入图像进行识别; (2)使用手写字体目标标识对步骤(1)中标识的节点进行文本内容标识,从而确定节点对应的特定原子; (3)将识别出的多个原子成对组合,然后使用化学键目标识别符对两个原子之间的化学键进行识别。 (4)在数据库中找到所识别原子的属性,计算结构式的相关属性,并将其输出;或将识别出的原子和原子之间的化学键存储为自定义的国王格式文件或绘制成新图像,然后输出。本发明解决了手绘图和图片中识别化学结构或反应式的问题,可广泛用于化学家的日常工作中。

著录项

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号