随着城市矿产资源循环利用技术的不断发展,废旧手机回收已成为当前研究热点。受限于计算资源和数据资源的相对缺乏,目前基于线下智能回收装备的废旧手机识别精度难以达到实际应用。针对上述问题,提出一种基于多元特征异构集成深度学习...随着城市矿产资源循环利用技术的不断发展,废旧手机回收已成为当前研究热点。受限于计算资源和数据资源的相对缺乏,目前基于线下智能回收装备的废旧手机识别精度难以达到实际应用。针对上述问题,提出一种基于多元特征异构集成深度学习的图像识别模型。首先,利用字符级文本检测算法(character region awareness for text detection,CRAFT)提取手机背部字符区域,再利用ImageNet预训练的VGG19模型作为图像特征嵌入模型,利用迁移学习理念提取待回收手机的局部字符特征和全局图像特征;然后,利用局部特征构建神经网络模式光学字符识别(optical character recognition,OCR)模型,利用全局和局部特征构建非神经网络模式深度森林分类(deep forest classification,DFC)模型;最后,将异构OCR和DFC识别模型输出的结果与向量组合后输入Softmax进行集成,基于权重向量得分最大准则获取最终识别结果。基于废旧手机回收装备的真实图像验证了所提方法的有效性。展开更多
引言 OCR(Optical Character Recognition,光学字符识别)属于一种高效的文字输入方式,亦可称之为文字识别。OCR技术的运用过程通常涉及将纸张上的文字、图像信息转化为计算机能识别的格式[1]。在档案工作“存量数字化、增量电子化”的...引言 OCR(Optical Character Recognition,光学字符识别)属于一种高效的文字输入方式,亦可称之为文字识别。OCR技术的运用过程通常涉及将纸张上的文字、图像信息转化为计算机能识别的格式[1]。在档案工作“存量数字化、增量电子化”的要求下,研究OCR识别在民生档案数字化管理中的应用,设计基于OCR识别的档案数字化管理方案,有助于解决纸质档案在扫描、识别、分类等环节容易出错且耗费大量人力的问题,提升民生服务效率,推动信息化建设再上新台阶。展开更多
Handwritten character recognition is considered challenging compared with machine-printed characters due to the different human writing styles.Arabic is morphologically rich,and its characters have a high similarity.T...Handwritten character recognition is considered challenging compared with machine-printed characters due to the different human writing styles.Arabic is morphologically rich,and its characters have a high similarity.The Arabic language includes 28 characters.Each character has up to four shapes according to its location in the word(at the beginning,middle,end,and isolated).This paper proposed 12 CNN architectures for recognizing handwritten Arabic characters.The proposed architectures were derived from the popular CNN architectures,such as VGG,ResNet,and Inception,to make them applicable to recognizing character-size images.The experimental results on three well-known datasets showed that the proposed architectures significantly enhanced the recognition rate compared to the baseline models.The experiments showed that data augmentation improved the models’accuracies on all tested datasets.The proposed model outperformed most of the existing approaches.The best achieved results were 93.05%,98.30%,and 96.88%on the HIJJA,AHCD,and AIA9K datasets.展开更多
在现代社会,发票作为一种重要的证明凭证,被广泛应用于各个行业。传统的发票录入和处理方式,依靠人工手动识别并输入信息,不仅效率低下、成本高昂,而且容易出现误差。光学字符识别(optical character recognition,OCR)技术和二维码识别...在现代社会,发票作为一种重要的证明凭证,被广泛应用于各个行业。传统的发票录入和处理方式,依靠人工手动识别并输入信息,不仅效率低下、成本高昂,而且容易出现误差。光学字符识别(optical character recognition,OCR)技术和二维码识别技术已经成为自动化发票录入和处理的重要手段。基于OCR和二维码技术,实现发票的自动识别,将识别的发票信息录入关系型数据库管理软件中方便后续处理,提高企业财务管理的信息化程度和准确性,为企业改进管理和提高效益提供技术支持。展开更多
Road transport is currently one of the most important sectors affecting sustainable development and the improvement of the population’s standard of living. In some sub-Saharan African countries, including Burundi, th...Road transport is currently one of the most important sectors affecting sustainable development and the improvement of the population’s standard of living. In some sub-Saharan African countries, including Burundi, the transport structure is vulnerable, under attack, or even damaged or destroyed. This is prompting decision-makers to look for every possible way to enable dynamic management of the road system, as well as the collection of tax revenues attributable to this sector. To reach this stage, we postulate that the introduction of the Intelligent Transport System (ITS) into the road tax and fee collection process would make a significant contribution (road safety, zero cash on silk Safety Officers, payment of a fine, eradication of road corruption etc.) to the digitization of the various transport sectors. As far as the city of Bujumbura is concerned (our field of intervention), the applicability of the present System could thus meet the expectations of the decision-maker, certain drivers and, by the same token, contribute to the promotion of Digital Technology in Burundi.展开更多
文摘随着城市矿产资源循环利用技术的不断发展,废旧手机回收已成为当前研究热点。受限于计算资源和数据资源的相对缺乏,目前基于线下智能回收装备的废旧手机识别精度难以达到实际应用。针对上述问题,提出一种基于多元特征异构集成深度学习的图像识别模型。首先,利用字符级文本检测算法(character region awareness for text detection,CRAFT)提取手机背部字符区域,再利用ImageNet预训练的VGG19模型作为图像特征嵌入模型,利用迁移学习理念提取待回收手机的局部字符特征和全局图像特征;然后,利用局部特征构建神经网络模式光学字符识别(optical character recognition,OCR)模型,利用全局和局部特征构建非神经网络模式深度森林分类(deep forest classification,DFC)模型;最后,将异构OCR和DFC识别模型输出的结果与向量组合后输入Softmax进行集成,基于权重向量得分最大准则获取最终识别结果。基于废旧手机回收装备的真实图像验证了所提方法的有效性。
文摘引言 OCR(Optical Character Recognition,光学字符识别)属于一种高效的文字输入方式,亦可称之为文字识别。OCR技术的运用过程通常涉及将纸张上的文字、图像信息转化为计算机能识别的格式[1]。在档案工作“存量数字化、增量电子化”的要求下,研究OCR识别在民生档案数字化管理中的应用,设计基于OCR识别的档案数字化管理方案,有助于解决纸质档案在扫描、识别、分类等环节容易出错且耗费大量人力的问题,提升民生服务效率,推动信息化建设再上新台阶。
文摘Handwritten character recognition is considered challenging compared with machine-printed characters due to the different human writing styles.Arabic is morphologically rich,and its characters have a high similarity.The Arabic language includes 28 characters.Each character has up to four shapes according to its location in the word(at the beginning,middle,end,and isolated).This paper proposed 12 CNN architectures for recognizing handwritten Arabic characters.The proposed architectures were derived from the popular CNN architectures,such as VGG,ResNet,and Inception,to make them applicable to recognizing character-size images.The experimental results on three well-known datasets showed that the proposed architectures significantly enhanced the recognition rate compared to the baseline models.The experiments showed that data augmentation improved the models’accuracies on all tested datasets.The proposed model outperformed most of the existing approaches.The best achieved results were 93.05%,98.30%,and 96.88%on the HIJJA,AHCD,and AIA9K datasets.
文摘在现代社会,发票作为一种重要的证明凭证,被广泛应用于各个行业。传统的发票录入和处理方式,依靠人工手动识别并输入信息,不仅效率低下、成本高昂,而且容易出现误差。光学字符识别(optical character recognition,OCR)技术和二维码识别技术已经成为自动化发票录入和处理的重要手段。基于OCR和二维码技术,实现发票的自动识别,将识别的发票信息录入关系型数据库管理软件中方便后续处理,提高企业财务管理的信息化程度和准确性,为企业改进管理和提高效益提供技术支持。
文摘Road transport is currently one of the most important sectors affecting sustainable development and the improvement of the population’s standard of living. In some sub-Saharan African countries, including Burundi, the transport structure is vulnerable, under attack, or even damaged or destroyed. This is prompting decision-makers to look for every possible way to enable dynamic management of the road system, as well as the collection of tax revenues attributable to this sector. To reach this stage, we postulate that the introduction of the Intelligent Transport System (ITS) into the road tax and fee collection process would make a significant contribution (road safety, zero cash on silk Safety Officers, payment of a fine, eradication of road corruption etc.) to the digitization of the various transport sectors. As far as the city of Bujumbura is concerned (our field of intervention), the applicability of the present System could thus meet the expectations of the decision-maker, certain drivers and, by the same token, contribute to the promotion of Digital Technology in Burundi.
