Thông tin chung
  English
  Đề tài NC khoa học
  Bài báo, báo cáo khoa học
  Hướng dẫn Sau đại học
  Sách và giáo trình
  Các học phần và môn giảng dạy
  Giải thưởng khoa học, Phát minh, sáng chế
  Khen thưởng
  Thông tin khác
  Tài liệu tham khảo
  Hiệu chỉnh
 
Số người truy cập: 107,419,385

 Improving the Efficiency of Image Retrieval Based on a Model Combining Graph and SOM
Tác giả hoặc Nhóm tác giả: Nguyen Thi Uyen Nhi, Van The Thanh, Nguyen Minh Hai
Nơi đăng: ICIC Express Letters Part B: Applications; Số: 14(5);Từ->đến trang: 457-466;Năm: 2023
Lĩnh vực: Công nghệ thông tin; Loại: Bài báo khoa học; Thể loại: Quốc tế
TÓM TẮT
Image retrieval is a research issue that has been of interest in recent times. In our previous work, we constructed a neighbor graph (called Graph-CTree) for storing and retrieving large image data. In this paper, we propose methods to improve the efficiency of image retrieval: creating a SgC Tree model from a combination of the neighbor graph (Graph-CTree) and self-organizing map (SOM). In this paper, SOM is assembled from clusters of Graph-CTree graphs, called grSOM, with input weight vectors taken during training C-Tree. So, the weights do not have to be tweaked too much during training, so the training time of grSOM is faster. grSOM is more flexible and allows scaling after training. Content-based image retrieval system has been built on SgC-Tree, called CBIRSgC, and experimented on COREL and ImageCLEF datasets to evaluate the effectiveness and correctness of the proposal.
ABSTRACT
Image retrieval is a research issue that has been of interest in recent times. In our previous work, we constructed a neighbor graph (called Graph-CTree) for storing and retrieving large image data. In this paper, we propose methods to improve the efficiency of image retrieval: creating a SgC Tree model from a combination of the neighbor graph (Graph-CTree) and self-organizing map (SOM). In this paper, SOM is assembled from clusters of Graph-CTree graphs, called grSOM, with input weight vectors taken during training C-Tree. So, the weights do not have to be tweaked too much during training, so the training time of grSOM is faster. grSOM is more flexible and allows scaling after training. Content-based image retrieval system has been built on SgC-Tree, called CBIRSgC, and experimented on COREL and ImageCLEF datasets to evaluate the effectiveness and correctness of the proposal.
[ icc_230425_172316.pdf ]
© Đại học Đà Nẵng
 
 
Địa chỉ: 41 Lê Duẩn Thành phố Đà Nẵng
Điện thoại: (84) 0236 3822 041 ; Email: dhdn@ac.udn.vn