This paper describes a quality-dependent score-level fusion framework of face, gait, and the height biomet- rics from a single walking image sequence. Individual person authentication accuracies by face, gait, and the height biometrics, are in general degraded when spatial resolution (image size) and temporal resolution (frame-rate) of the input image sequence decrease and the degree of such accuracy degradation differs among the individual modalities. We therefore set the optimal weights of the individual modalities based on linear logistic regression framework de- pending on a pair of the spatial and temporal resolutions, which are called qualities in this paper. On the other hand, it is not a realistic solution to compute and store the optimal weights for all the possible qualities in advance, and also the optimal weights change across the qualities in a nonlinear way. We thus propose a method to estimate the optimal weights for arbitrary qualities from a limited training pairs of the optimal weights and the qualities, based on Gaussian process regression with a nonlinear kernel function. Experiments using a publicly available large population gait database with 1,935 subjects under various qualities, showed that the person authentication accuracy improved by successfully estimating the weights depending on the qualities.

This paper presents a novel method for detecting 3D road boundaries, such as walls, guardrails, and curbs, using on-board stereo cameras. The proposed method uses conformal geometric algebra, which can describe different shapes in a common representation. 3D road boundaries on straight and curved roads are seamlessly detected by use of this representation, and this framework is also applied to curb detection by a subtle modification. Experimental results show that despite its algorithmic simplicity, the proposed method exhibited competitive detection performance compared with conventional model fitting and curb detection methods.

This paper discusses about object detection based on spatio-temporal light field sensing. Our proposed method generates an arbitrary in-focus plane in the surveillance scene, and the background region can be filtered out by out-focusing. A new feature representation, called Local Ray Pattern (LRP), is introduced to evaluate the spatial consistency of light rays. The combination of LRP and GMM-based background modeling realizes object detection on the in-focus plane. Experimental results demonstrate the effectiveness and applicability for video surveillance.

Image recognition in client server system has a problem of data traffic. However, reducing data traffi cg ives rise to worsening of performance. Therefore, we represent binary codes as high dimensional local features in client side, and represent real vectors in server side. As a result, we can suppress the worsening of the performance, but it problems of an increase in the computational cost of the distance computation and a different scale of norm between feature vectors. Therefore, to solve the first problem, we optimize the scale factor so as to absorb the scale difference of Euclidean norm. For second problem, we compute efficiently the Euclidean distance by decomposing the real vector into weight factors and binary basis vectors. As a result, the proposed method achieves the keypoint matching with high-speed and high-precision even if the data traffic was reduced.

The best known method for optimally computing parameters from noisy data based on geometric constraints is maximum likelihood (ML). This paper reinvestigates “hyperaccurate correction” for further improving the accuracy of ML. In the past, only the case of a single scalar constraint was studied. In this paper, we extend it to multiple constraints given in the form of vector equations. By detailed error analysis, we illuminate the existence of a term that has been ignored in the past. Doing simulation experiments of ellipse fitting, fundamental matrix, and homography computation, we show that the new term does not effectively affect the final solution. However, we show that our hyperaccurate correction is even superior to hyper-renormalization, the latest method regarded as the best fitting method, but that the iterations of ML computation do not necessarily converge in the presence of large noise.

In this paper, we propose a new approach for recognizing group events and abnormality detection in a crowded scene. A manifold learning algorithm with temporal-constraints is proposed to embed a video of a crowded scene in a low-dimensional space. Our low dimensional representation of a video preserves the spatial temporal property of a video as well as the characteristic of the video. Recognizing video events and abnormality detection in a crowded scene is achieved by studying the video trajectory in the manifold space. We evaluate our proposed method on the state-of-the-art public data-sets containing different crowd events. Qualitative and quantitative results show the promising performance of the proposed method.

In this paper, an adaptive framework based on histogram separation and mapping for image contrast enhancement is presented. In this framework, the histogram is separated by binary tree structure with the proposed adaptive histogram separation strategy. Generally, histogram equalization (HE) is an effective technique for contrast enhancement. However, the conventional HE usually gives the processed image with unnatural look and artifacts by excessive enhancement. For overcoming this shortage, the adaptive histogram separation unit (AHSU) is proposed to convert the global enhancement problem into local. And for mapping the histogram partitions into more optimal ranges, the exact histogram separation is discussed. Finally, an adaptive histogram separation and mapping framework (AHSMF) for contrast enhancement is presented, and the experimental results show better effectiveness than other histogram based methods.

In this paper we present new efficient solutions to the absolute pose problems for cameras with unknown focal length and unknown focal length and radial distortion from four 2D-to-3D point correspondences. We propose to solve these problems separately for non-planar and for planar scenes. By decomposing the problems into these two situations we obtain simpler and more efficient solvers than the previously known general solvers. We demonstrate in synthetic and real experiments significant speedup of our solvers. Especially our new solvers for absolute pose problem for camera with unknown focal length and radial distortion are about 40 × (non-planar) and 160 × (planar) faster than the general solver. Moreover, we show that our specific solvers can be joined into new general solvers, based on performing either planar or non-planar solver according to the scene structure or performing both solvers simultaneously and selecting the better result. Such joined solvers give comparable or even better results than the existing general solvers for planar as well as non-planar scenes.

The importance of person identification techniques is increasing for visual surveillance applications. In social living scenarios, people often act in groups composed of friends, family, and co-workers, and this is a useful cue for person identification. This paper describes a method for person identification in video sequences based on this group cue. In the proposed approach, the relationships between the people in an input sequence are modeled using a graphical model. The identity of each person is then propagated to their neighbors in the form of message passing in a graph via belief propagation, depending on each person's group affiliation information and their characteristics, such as spatial distance and velocity vector difference, so that the members of the same group with similar character- istics enhance each other's identities as group members. The proposed method is evaluated through gait-based person identification experiments using both simulated and real input sequences. Experimental results show that the identi- fication performance is considerably improved when compared with that of the straightforward method based on the gait feature alone.

We analyzed a light-field super-resolution problem in which the 3-D scene is reconstructed with a higher resolution using super-resolution (SR) reconstruction with a given set of multi-view images with a low resolution. The arrangement of the multi-view cameras is important because it determines the quality of the reconstruction. To simplify the analysis, we considered a situation in which a plane is located at a certain depth and a texture on that plane is super-resolved. We formulated the SR reconstruction process in the frequency domain, where the camera arrangement can be independently expressed as a matrix in the image formation model. We then evaluated the condition number of the matrix to quantify the quality of the SR reconstruction. We clarified that when the cameras are arranged in a regular grid, there exist singular depths in which the SR reconstruction becomes ill-posed. We also determined that this singularity can be avoided if the arrangement is randomly perturbed.