Unicursal Gesture Recognition with Support Vector Machines Using Distance to Borders

スポンサーリンク

概要

• 論文の詳細を見る

• Vision-based recognition systems using hand gestures to control applications like TV menus are gaining popularity in the market. To improve gesture recognition, we propose a combination of Support Vector Machines (SVM) and Distance to Border features (DtB) to recognize unicursal shapes drawn by finger motion or hand shapes such as peace sign. Two SVM models are trained and combined to classify each gesture. Our algorithm is simple, easily implemented but proper image processing for hand gestures is required before classification using the algorithm. The performance of our proposed recognition algorithm is evaluated on eight unicursal shapes and results are reported.

• 社団法人電子情報通信学会の論文
• 2011-10-06

著者

• KAGAMI Shingo

  Tohoku University

• Kobayashi Minoru

  Ntt Cyber Solutions Laboratories Ntt Corporation

• Aoki Ryosuke

  Ntt Cyber Solutions Laboratories Ntt Corporation:tohoku University

• Kobayashi Toru

  Ntt Cyber Solutions Laboratories Ntt Corporation

• CHAN Bessie

  NTT Cyber Solutions Laboratories, NTT Corporation

• IHARA Masayuki

  NTT Cyber Solutions Laboratories, NTT Corporation

• Chan Bessie

  Ntt Cyber Solutions Laboratories Ntt Corporation

• Ihara Masayuki

  Ntt Cyber Solutions Laboratories Ntt Corporation

関連論文

• 2P2-B24 Illumination-based Synchronization of High-Speed Vision Sensors
• Unicursal Gesture Recognition with Support Vector Machines Using Distance to Borders

スポンサーリンク