닫기
Loading..

전자정보연구정보센터 ICT 융합 전문연구정보의 집대성

학술대회 프로시딩

홈 홈 > 연구문헌 > 학술대회 프로시딩 > 대한전자공학회 학술대회 > 2019년 대한전자공학회 하계종합학술대회

2019년 대한전자공학회 하계종합학술대회

Current Result Document : 189 / 575 이전건 이전건   다음건 다음건

한글제목(Korean Title) 한국 수화 인식을 위한 데이터셋
영문제목(English Title)
저자(Author) 양승한   정승준   강희광   김창익  
원문수록처(Citation) VOL 42 NO. 01 PP. 0480 ~ 0488 (2019. 06)
한글내용
(Korean Abstract)
영문내용
(English Abstract)
Recently, the development of computer vision technologies has also shown excellent performance in complex tasks such as behavioral recognition. Therefore, several studies provide datasets for behavior recognition tasks, including sign language datasets. In many countries, researchers are already carring out studies to automatically recognize and interpret sign language to facilitate communication with deaf people. However, there is no dataset aiming at sign language recoggnition that is used in Korea yet, and research on this is insufficient. Since sign language varies from country to country, the need for a dataset for Korean sign language is sufficient. Therefore, this paper aims to provide a dataset that can incorporate Korean sign language into behavior recognition technology using deep learning. The existing Korean sign language video is distributed for educational purposes and is not suitable for network learning. So we present the Korean Sign Language (KSL)dataset. The dataset was composed of sign language by 20 deaf people, all of which consist of 77 words. We train and evaluate this dataset in deep learning networks that have recently achieved excellent performance in the behavior recognition task. Also, we have confirmed through the deconvolution-based visualization method that the deep learning network fully understands the characteristics of the dataset.
키워드(Keyword)   
원문 PDF 다운로드