講演抄録/キーワード |
講演名 |
2018-12-14 11:00
[ショートペーパー]Skeleton-based Human Action Recognition with Fine-to-Coarse Convolutional Neural Network Thao Minh Le・○Nakamasa Inoue・Koichi Shinoda(TokyoTech) PRMU2018-86 |
抄録 |
(和) |
(まだ登録されていません) |
(英) |
This work introduces a new framework for skeleton-based human action recognition. Existing approaches using Convolutional Neural Network (CNN) often suffer from the insufficiency problem of training data. In this study, we utilize a fine-to-coarse (F2C) CNN architecture that is come up based on the special structure of human skeletal data. We evaluate our proposed method on two skeletal datasets publicly available, namely NTU RGB+D and SBU Kinect Interaction dataset. It achieves 79.6% and 84.6% of accuracies on NTU RGB+D with cross-object and cross-view protocol, respectively, which are almost identical with the state-of-the-art performance. In addition, our method significantly improves the accuracy of the actions in two-person interactions. |
キーワード |
(和) |
/ / / / / / / |
(英) |
Action Recognition / Deep Learning / / / / / / |
文献情報 |
信学技報, vol. 118, no. 362, PRMU2018-86, pp. 61-64, 2018年12月. |
資料番号 |
PRMU2018-86 |
発行日 |
2018-12-06 (PRMU) |
ISSN |
Online edition: ISSN 2432-6380 |
著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します.(許諾番号:10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
PDFダウンロード |
PRMU2018-86 |