手の甲の皮膚変形計測による手形状の三次元再構成
3D Reconstruction of Hand Postures by Measuring Skin Deformation on Back Hand
2017
久能若葉，杉浦裕太，浅野直生，河合航，杉本麻樹
Wakaba Kuno, Yuta Sugiura, Nao Asano, Maki Sugimoto
[Reference /引用はこちら]
Wakaba Kuno, Yuta Sugiura, Nao Asano, Wataru Kawai, and Maki Sugimoto, 3D Reconstruction of Hand Postures by Measuring Skin Deformation on Back Hand, In Proceedings of the International Conference on Artificial Reality and Telexistence & Eurographics Symposium on Virtual Environments (ICAT-EGVE ’17), 8 pages. [DOI]
3D Reconstruction of Hand Postures by Measuring Skin Deformation on Back Hand (ICAT-EGVE 2017) from sugiuralab
本研究では装着型デバイスを用いて手の甲の変形から指の姿勢を推定する手法を提案する。本手法は回帰分析を利用しており、2つの連続値変数の相関関係を表す回帰モデルを手の甲データと指姿勢データによって学習することで手の甲の変形から指の姿勢を推定する。指の位置を直接計測せず手の甲の変形を計測するので手指の自然な動きを阻害することなく指の姿勢を推定できる。本研究は実環境のユーザの指姿勢を利用したバーチャル環境とのインタラクションに利用できる。提案手法の評価のためにユーザ実験を行い、静止状態と動作状態での手の甲の変形と指の姿勢を計測し指姿勢推定誤差による評価を行った。
We propose a method for reconstructing hand posture by measuring the deformation of the back of the hand with a wearable device. The deformation of skin on the back of the hand can be measured by using several photo-reflective sensors attached to a wearable device. In the learning phase, our method constructs a regression model by using the data on hand posture captured by a depth camera and data on the skin deformation of the back of the hand captured by several photo sensors. In the estimation phase, by using this regression model, the posture of the hand is reconstructed from the data of the photo sensors in real-time. The posture of fingers can be estimated without hindering the natural movement of the fingers since the deformation of the back of the hand is measured without directly measuring the position of the fingers. This method can be used by users to manipulate information in a virtual environment with their fingers. We conducted an experiment to evaluate the accuracy of reconstructing hand posture with the proposed system.