초음파 배열 시스템 기반의 CNN을 활용한 손짓 인식 정확도 향상

다국어 초록 (Multilingual Abstract)

The aim of this paper is to develop a device that provides high accuracy in the field of hand gesture recognition by utilizing ultrasound transducer arrays and CNN (Convolutional Neural Network) technology. Ultrasound technology has been continuously researched in the past, and its advantages such as high resolution and low cost make it highly applicable in the field of hand gesture recognition. Additionally, CNN has shown high performance in the field of image processing, enabling the creation of a model that analyzes and classifies hand gesture data. In this study, these technologies were combined to improve the accuracy of hand gesture recognition. Five sign language gesture data were collected using an 7x1 ultrasound transducer array system, and the accuracy was improved by inputting the preprocessed training data into CNN. The proposed model achieved an accuracy of approximately 91%.

목차 (Table of Contents)

• Abstract
• Ⅰ. 서론
• Ⅱ. 이론
• Ⅲ. 실험 및 딥러닝 학습
• Ⅳ. 결론 및 향후 연구 방향
• 참고문헌