歩き方から個人識別を行う.従来手法は,画像(または画像特徴量)マッチングを基礎とするため,カメラ視点に依存する.テストデータと訓練データのカメラ視点が大きく異なる場合,その識別精度は低くなる.我々はカメラ視点に依存しない識別手法を提案する.歩行動作の位相およびカメラ視点を状態変数とする状態空間モデルCyclic motion modelを用いる.識別候補者それぞれのモデルパラメータを学習し,歩行を表現する.Particle Filterを用いて観測画像列に対応する状態変数を推定する.そして,モデル尤度を用いて識別を行う. / We challenge the human identification problem from gait and body shape. Conventional methods depend on the camera viewing direction since they are based on matching images, their accuracy is low when there is a big difference between the camera viewing direction of the test and training. We propose a novel method that does not depend on the camera viewing direction. We develop a state space model called "cyclic motion model" whose state variables are the phase and the camera viewing direction. We learn model parameters for each candidate person, and represent their walking. To identify a person from the observed image sequence, we first compute the model likelihoods using a particle filter. We then identify the person from the model likelihoods.