Current gait recognition approaches only consider individuals walking frontoparallel to the image plane. This makes them inapplicable for recognizing individuals walking from different angles with respect to the image plane. In this paper, we propose a kinematic-based approach to recognize individuals by gait. The proposed approach eastimates 3D human walking parameters by performing a least squares fit of the 3D kinematic model to the 2D silhouette extracted from a monocular image sequence. A genetic algorithm is used for feature selection from the estimated parameters, and the individuals are then recognized from the feature vectors using a nearest neighbor method. Experimental results show that the proposed approach achieves good performance in recognizing individuals walking from different angles with respect to the image plane.