Based on personal information protection act (PIPA), collecting and distribution of the personal information are strictly prohibited, therefore the collecting the personal information is nearly impossible. To solve this problem, this study propose synthetic datasets based optical character recognition (OCR). In this study, driver’s license is chosen as for collecting personal information because the driver’s license contains detailed personal information and therefore it is very hard to obtain the information in real world. To realize the proposed method, generative adversarial network (GAN) based driver’s licenses are synthesized, and lightweight convolutional recurrent network-connectionist temporal classification (CRNN-CTC) based optical character reconition (OCR) is proposed. Experimental results shows a character error rate of 0.01 and a sentence error rate of 0.16. Qualitative experimental result also demonstrates the effectiveness of the proposed method.
한국어
개인정보 보호법에 의하여, 개인 정보의 수집 및 배포 등은 엄격히 제한을 받고 있어서, 개인정보의 수집 은 대단히 어렵다. 이러한 개인정보 수집의 어려움을 해결하고자 합성 데이터 기반의 광학 문자 인식(Optical Character Recognition, OCR) 시스템을 제안한다. 이 연구에서는 개인정보를 얻는 대상은 운전면허증으로 선택하 였는데, 그 이유는 여러 가지 개인정보가 들어있고 개인정보 보호로 인하여 실제 데이터를 얻기 어렵기 때문이다. 이를 구현하기 위하여 생성적 적대신경망(Generative Adversarial Network, GAN) 기반의 운전면허증을 합성하 여 데이터세트을 구축하였고, 문자 인식기로는 휴대가 가능한 경량 컨볼루셔널 리커런트 신경망-연결주의 시계열 분류(Convolutional Recurrent Network-Connectionist Temporal Classification, CRNN-CTC)를 구현하였다. 실 험을 통하여, 문자오차율과 문장 오차율은 각각 0.2778%와 2.4167%을 얻었고, 정성 실험을 진행하여 제안하는 방 법이 효과적임을 증명하였다.
목차
요약 Abstract Ⅰ. 서론 Ⅱ. 관련연구 Ⅲ. 광학 문자 판독기 설계 3.1 데이터세트 설계 3.2 광학 문자 판독기 설계 3.3 손실 함수 3.4 성능 지표 Ⅳ. 실험 결과 Ⅴ. 결론 REFERENCES
Ever since next generation convergence technology became one of the most important industries in the nation, computing professionals have encountered a growing number of challenges. Along with scholars and colleagues in related fields, they have gathered in avariety of forums and meetings over the last few decades to share their knowledge, experiences and the outcome of their research. These exchanges have led to the founding of the International Next-generation Convergence technology (INCA) on December 1, 2015. INCA was registered as an incorporated association under the Ministry of Information and Communications. The main purpose of the organization is to improve our society by achieving the highest capability possible in next generation convergence technology.
간행물
간행물명
차세대융합기술학회논문지 [The Journal of Next-generation Convergence Technology Association]