| 제목 | 추가학습이 불필요한 이미지 특징 유사도 기반상품 식별 시스템 |
|---|---|
| Title | Product Identification System based on Image FeatureSimilarity with Learning-free Model |
| 저자 |
유영재 (서울대학교 컴퓨터공학부), (투모로 로보틱스) 윤혜정 (서울대학교 협동과정 인공지능전공) 김준오 (서울대학교 협동과정 인공지능전공) 박예솔 (서울대학교 협동과정 인공지능전공) 장병탁* (서울대학교 컴퓨터공학부), (서울대학교 협동과정 인공지능전공), (투모로 로보틱스) |
| Author |
Youngjae Yoo(Department of Computer Science and Engineering, Seoul National University),(Tommoro Robotics) HyeJung Yoon(Interdisciplinary Program in Artificial Intelligence, Seoul National University) Juno Kim(Interdisciplinary Program in Artificial Intelligence, Seoul National University) Yesol Park(Interdisciplinary Program in Artificial Intelligence, Seoul National University) Byoung-Tak Zhang†(Department of Computer Science and Engineering, Seoul National University),(Interdisciplinary Program in Artificial Intelligence, Seoul National University),(Tommoro Robotics) |
| Bibliography | Journal of Logistics Science & Technology, 3(2),36~55, 2022, |
| DOI | |
| Key Words | Object Recognition, Image Feature Matching, Learning-free Model, Logistics Automation |
| Abstract | Recently, logistics centers attempts to get help from artificial intelligence robots with hard labor. To generally perform picking task for robots, it is essential to detect and identify the product through the camera. Supervised learning-based deep learning technology is suitable for recognizing objects with high accuracy. But it has a disadvantage that requires a lot of time because a human must manually label the answer of the train image. In this paper, we propose a method that minimizes manual labor and makes it easy to add products. Our algorithm identifies the product with the highest similarity by calculating the similarity between the features of the input image and the features of the images in the product database. It does not require learning and labeling to identify a new product. To verify it, we test the algorithm by photographing a test product images in an environment that simulates a logistics site. |
| PDF download | JM_3[2]-P36~55.pdf |