In this paper we deal with the problem of calibrating a rotating and zooming camera, without 3D pattern, whose internal calibration parameters change frame by frame. First, we theoretically show the existence of the calibration parameters up to an orthogonal transformation under the assumption that the skew of the camera is zero. Auto-calibration becomes possible by analyzing inter-image homographies which can be obtained from the matches in images of the same scene, or through direct nonlinear iteration. In general, at least four homographies are needed for auto-calibration. When we further assume that the aspect ratio is known and the principal point is fixed during the sequence then one homography yields camera parameters, and when the aspect ratio is assumed to be unknown with fixed principal point then two homographies are enough. In the case of a fixed principal point, we suggest a method for obtaining the calibration parameters by searching the space of the principal point. If this is not the case, then nonlinear iteration is applied. The algorithm is implemented and validated on several sets of synthetic data. Also experimental results for real images are given.

We propose a class-incremental learning framework for human activity recognition based on the Bag-of-Sequencelets model (BoS). The framework updates learned models efficiently without having to relearn them when training data of new classes are added. In this framework, all types of features including hand-crafted features and Convolutional Neural Networks (CNNs) based features and combinations of those features can be used as features for videos. Compared with the original BoS, the new framework can reduce the learning time greatly with little loss of classification accuracy.

This paper deals with the problem of multiple object tracking with the condensation algorithm, applied to tracking of soccer players. To solve the problem of failures in tracking multiple players under overlapping, we introduce occlusion alarm probability, which attracts or repels particles based on their posterior distribution of previous time step. Real experiments showed a robust performance.

An algorithm is developed for augmenting a real video with virtual graphics objects without computing Euclidean information. For this, we design a method of specifying the virtual camera that performs Euclidean orthographic projection in recovered affine space. In addition, our method has the capability of generating views of objects shaded by virtual light sources. Our novel formulation and experimental results are presented.

This paper proposes a linear algorithm for metric reconstruction from projective reconstruction. Metric reconstruction problem is equivalent to estimating the projective transformation matrix that converts projective reconstruction to Euclidean reconstruction. We build a quadratic form from dual absolute conic projection equation with respect to the elements of the transformation matrix. The matrix of quadratic form of rank 2 is then eigen-decomposed to produce a linear estimate. The algorithm is applied to three different sets of real data and the results show a feasibility of the algorithm. Additionally, our comparison of results of the linear algorithm to results of bundle adjustment, applied to sets of synthetic image data having Gaussian image noise, shows reasonable error ranges.