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: 109,882,814

 Video-based Action Recognition to Assist Mild Cognitive Impairment Prediction
Tác giả hoặc Nhóm tác giả: Hoang Le Uyen Thuc, Shian-Ru Ke and Pham Van Tuan
Nơi đăng: International Journal of Computer Science and Information Security, ISSN 1947 5500 (ESCI); Số: vol. 14(10);Từ->đến trang: 24-33;Năm: 2016
Lĩnh vực: Công nghệ thông tin; Loại: Bài báo khoa học; Thể loại: Quốc tế
TÓM TẮT
In this paper, we propose to use video-based human action recognition as an effective solution to automatically detect abnormal actions that may indicate mild cognitive impairment (MCI) from monocular action video sequences. There are two main modules in the proposed system. The first module is based on our previous action recognition technique which combines 3D geometric relational features (3D GFR) and hidden Markov model (HMM). First, we analyze each video frame to estimate 3D coordinates of 13 specific body points, then convert these 3D coordinates into corresponding 3D GRF vectors which describe the geometric relations between interested body parts. Second, we recognize different human actions by using HMM whose parameters are appropriately chosen via experiments. The recognition results are further analyzed in the second module to detect occurred abnormal actions which can be used to predict MCI. The experimental results on IXMAS database show the effectiveness of the solution in terms of promising abnormal action pattern detection rates.
ABSTRACT
In this paper, we propose to use video-based human action recognition as an effective solution to automatically detect abnormal actions that may indicate mild cognitive impairment (MCI) from monocular action video sequences. There are two main modules in the proposed system. The first module is based on our previous action recognition technique which combines 3D geometric relational features (3D GFR) and hidden Markov model (HMM). First, we analyze each video frame to estimate 3D coordinates of 13 specific body points, then convert these 3D coordinates into corresponding 3D GRF vectors which describe the geometric relations between interested body parts. Second, we recognize different human actions by using HMM whose parameters are appropriately chosen via experiments. The recognition results are further analyzed in the second module to detect occurred abnormal actions which can be used to predict MCI. The experimental results on IXMAS database show the effectiveness of the solution in terms of promising abnormal action pattern detection rates.
© Đại học Đà Nẵng
 
 
Địa chỉ: 41 Lê Duẩn Thành phố Đà Nẵng
Điện thoại: (84) 0236 3822 041 ; Email: dhdn@ac.udn.vn