This paper presents a novel technique for recovering 3D non-rigid shape models from 2D image sequences, particularly useful for applications such as facial animation and tracking. The method is based on a non-rigid model where the 3D shape in each frame is a linear combination of a set of basis shapes. Unlike existing techniques that assume rigidity, this approach allows for the recovery of complex, deformable objects. The tracking matrix, which represents the pose, configuration, and shape, is of higher rank and can be factored using singular value decomposition (SVD) in a three-step process. The authors demonstrate the effectiveness of their method on video sequences of people and animals, achieving high accuracy in recovering 3D non-rigid facial models. The technique does not require an a-priori model and can handle challenging scenarios with limited out-of-plane motion. Future work includes extending the method to handle occluded feature tracks and larger datasets.This paper presents a novel technique for recovering 3D non-rigid shape models from 2D image sequences, particularly useful for applications such as facial animation and tracking. The method is based on a non-rigid model where the 3D shape in each frame is a linear combination of a set of basis shapes. Unlike existing techniques that assume rigidity, this approach allows for the recovery of complex, deformable objects. The tracking matrix, which represents the pose, configuration, and shape, is of higher rank and can be factored using singular value decomposition (SVD) in a three-step process. The authors demonstrate the effectiveness of their method on video sequences of people and animals, achieving high accuracy in recovering 3D non-rigid facial models. The technique does not require an a-priori model and can handle challenging scenarios with limited out-of-plane motion. Future work includes extending the method to handle occluded feature tracks and larger datasets.