Recently, the team led by Associate Professor Guo Yang from the Department of Mathematics, College of Sciences, NEU has made important progress in the field of PnP problems in computer vision. Their research findings, “A Minimal Solution to the Perspective-3-Point Problem for the Camera with Unknown Focal Length and Two Degrees-of-Freedom Rotation,” have been accepted for publication in the top international academic journal in artificial intelligence and computer science, IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI). Associate Professor Guo Yang serves as both the first author and corresponding author, and Lu Xiaoxue, a 2022 master’s student of the Department of Mathematics, is the second author.
For the absolute pose estimation problem of cameras with unknown focal length and two degrees-of-freedom rotation, the paper proposes a minimal solution based on three perspective point correspondences. According to the imaging principle, the problem is transformed into four polynomial algebraic equations with four unknowns. By using the classical resultant elimination method and matrix operations, the problem is further converted into solving a univariate sixth-order polynomial equation. The paper proves that the upper bound of the number of physically meaningful solutions for this type of P3P problem is 6, and it provides an example demonstrating that this upper bound is attainable. The paper also presents the geometric configurations when the upper bound of the number of solutions is 2, 4, and infinite. Extensive simulations and real experiments show that the algorithm proposed in the paper, when combined with the Random Sample Consensus (RANSAC) algorithm, exhibits excellent robustness performance in absolute pose estimation. This paper provides innovative contributions with practical application value for the improvement and development of theoretical research in the field of PnP problems.
It is reported that TPAMI is widely recognized as a top international academic journal in the fields of artificial intelligence and computer science. It is a Class A journal in artificial intelligence recommended by the China Computer Federation (CCF), with an impact factor of 18.6, and it enjoys extremely high academic influence and academic recognition worldwide.
