Experimental results from simulated imagery as well as several real image sequences are included. For the simulated imagery, the bias and the variance of the estimators are studied by the Monte Carlo method; while for the real imagery, the ground truth of the motion and structure parameters (when available) is compared with the computed results."
The fully automated estimation of the 6 degrees of freedom camera motion and the imaged 3D scenario using as the only input the pictures taken by the camera has been a long term aim in the computer vision community. The associated line of research has been known as Structure from Motion (SfM). An intense research effort during the latest decades has produced spectacular advances; the topic has reached a consistent state of maturity and most of its aspects are well known nowadays. 3D vision has immediate applications in many and diverse fields like robotics, videogames and augmented reality; and technological transfer is starting to be a reality. This book describes one of the first systems for sparse point-based 3D reconstruction and egomotion estimation from an image sequence; able to run in real-time at video frame rate and assuming quite weak prior knowledge about camera calibration, motion or scene. Its chapters unify the current perspectives of the robotics and computer vision communities on the 3D vision topic: As usual in robotics sensing, the explicit estimation and propagation of the uncertainty hold a central role in the sequential video processing and is shown to boost the efficiency and performance of the 3D estimation. On the other hand, some of the most relevant topics discussed in SfM by the computer vision scientists are addressed under this probabilistic filtering scheme; namely projective models, spurious rejection, model selection and self-calibration.
This volume of original papers has been assembled to honor the achievements of Professor Thomas S Huang in the area of image processing and image analysis. Professor Huang's life of inquiry has spanned a number of decades as his work on imaging problems began in 1960's. Over these 40 years, he has made many fundamental and pioneering contributions to nearly every area of this field. Professor Huang has received numerous Awards, including the prestigious Jack Kilby Signal Processing Medal from IEEE. He has been elected to the National Academy of Engineering, and named Fellow of IEEE, Fellow of OSA, Fellow of IAPR, and Fellow of SPIE. Professor Huang has made fundamental contributions to image processing, pattern recognition, and computer vision: including design and stability test of multidimensional digital filters, digital holography; compression techniques for documents and images; 3D motion and modeling, analysis and visualization of the human face, hand and body, multi-modal human-computer interfaces; and multimedia databases. Many of his research ideas have been seminal, opening up new areas of research. Professor Huang is continuing his contribution to the field in the new millennium!This book is intended to highlight his contributions by showing the breadth of areas in which his students are working. As such, contributed chapters were written by some of his many former graduate students (some with Professor Huang as a coauthor) and illustrate not only his contributions to imaging science but also his commitment to educational endeavor. The breadth of contributions is an indication of influence of Professor Huang to the field of signal processing, image processing, computer vision and applications; the book includes chapters on learning in image retrieval, facial motion analysis, cloud motion tracking, wavelet coding, robust video transmission, and many other topics. The Appendix contains several reprints of Professor Huang's most influential papers from 1970's to 1990's. This book is directed towards image processing researchers, including academic faculty, graduate students and industry researchers, as well as toward professionals working in application areas.
"The main theme of the 1988 workshop, the 18th in this DARPA sponsored series of meetings on Image Understanding and Computer Vision, is to cover new vision techniques in prototype vision systems for manufacturing, navigation, cartography, and photointerpretation." P. v.
Motion and Structure from Image Sequences is invaluable reading for researchers, graduate students, and practicing engineers dealing with computer vision. It presents a balanced treatment of the theoretical and practical issues, including very recent results - some of which are published here for the first time. The topics covered in detail are: - image matching and optical flow computation - structure from stereo - structure from motion - motion estimation - integration of multiple views - motion modeling and prediction Aspects such as uniqueness of the solution, degeneracy conditions, error analysis, stability, optimality, and robustness are also investigated. These details together with the fact that the algorithms are accessible without necessarily studying the rest of the material, make this book particularly attractive to practitioners.
The fully automated estimation of the 6 degrees of freedom camera motion and the imaged 3D scenario using as the only input the pictures taken by the camera has been a long term aim in the computer vision community. The associated line of research has been known as Structure from Motion (SfM). An intense research effort during the latest decades has produced spectacular advances; the topic has reached a consistent state of maturity and most of its aspects are well known nowadays. 3D vision has immediate applications in many and diverse fields like robotics, videogames and augmented reality; and technological transfer is starting to be a reality. This book describes one of the first systems for sparse point-based 3D reconstruction and egomotion estimation from an image sequence; able to run in real-time at video frame rate and assuming quite weak prior knowledge about camera calibration, motion or scene. Its chapters unify the current perspectives of the robotics and computer vision communities on the 3D vision topic: As usual in robotics sensing, the explicit estimation and propagation of the uncertainty hold a central role in the sequential video processing and is shown to boost the efficiency and performance of the 3D estimation. On the other hand, some of the most relevant topics discussed in SfM by the computer vision scientists are addressed under this probabilistic filtering scheme; namely projective models, spurious rejection, model selection and self-calibration.
