The paper by S. S. Beauchemin and J. L. Barron from the University of Western Ontario discusses the computation of optical flow, which is the projection of three-dimensional object motion onto the image plane. Optical flow can be used to estimate 3D scene properties, motion parameters, and perform various tasks such as motion detection, object segmentation, and stereo disparity measurement. The authors classify and examine widely known methods for estimating optical flow, focusing on the hypotheses and assumptions they use. They also discuss current research issues and challenges, including the aperture problem, occlusion, transparency, and the need for reliable and accurate measurements. The paper covers different classes of optical flow techniques, such as intensity-based differential methods, multiconstraint methods, frequency-based methods, correlation-based methods, multiple motion methods, and temporal refinement methods. Each class is described in detail, with examples of representative state-of-the-art approaches. The authors highlight the importance of hierarchical processing to handle large image motions and the need for confidence measures to indicate the reliability of computed velocities. They also discuss the limitations and improvements in optical flow computation, emphasizing the accuracy, density, and reliability of measurements.The paper by S. S. Beauchemin and J. L. Barron from the University of Western Ontario discusses the computation of optical flow, which is the projection of three-dimensional object motion onto the image plane. Optical flow can be used to estimate 3D scene properties, motion parameters, and perform various tasks such as motion detection, object segmentation, and stereo disparity measurement. The authors classify and examine widely known methods for estimating optical flow, focusing on the hypotheses and assumptions they use. They also discuss current research issues and challenges, including the aperture problem, occlusion, transparency, and the need for reliable and accurate measurements. The paper covers different classes of optical flow techniques, such as intensity-based differential methods, multiconstraint methods, frequency-based methods, correlation-based methods, multiple motion methods, and temporal refinement methods. Each class is described in detail, with examples of representative state-of-the-art approaches. The authors highlight the importance of hierarchical processing to handle large image motions and the need for confidence measures to indicate the reliability of computed velocities. They also discuss the limitations and improvements in optical flow computation, emphasizing the accuracy, density, and reliability of measurements.