Search

• The curved LCD is first applied to the PMD, which increases the measurable curvature than the planar LCD. • A calibration method is proposed for the system based on the 3D fiducials on the curved LCD rather than the 2D fiducials on the planar one. • Different from the traditional stereo-PMD in which only areas that has common FOV with phase information of two cameras can be reconstructed, the proposed stereo-PMD reconstruction and registration algorithm can reconstruct the surface in FOV of each camera, respectively, and then register together. • The reconstructed error of RMS is less than 100 nm for the concave mirror of ROC 1000 mm. Phase measuring deflectometry (PMD) technique is widely used to measure the specular surface. However, the Current PMD methods are generally based on the planar LCD for the measurement, which limits the measurable curvature of the surface under test (SUT). To increase the measurable range of curvature, this paper proposes a novel PMD method by using a curved LCD monitor. A calibration method based on the cylindrical coordinate system is introduced for the proposed system calibration. To extend the measurable range further, the curved LCD is applied in stereoscopic PMD. The partial surface of SUT in the field of view (FOV) of each camera and curved LCD are reconstructed with iterative algorithm respectively, the calibration parameters as well as the seed point are used for point cloud registration. Simulations show the advantage of a curved LCD over a planar one and the feasibility of the proposed method. The actual performance is demonstrated in the experiments for specular optics and a phone touchscreen. [ABSTRACT FROM AUTHOR]

Abstract: It has been found that facial thermograms vary with ambient temperature, as well as other internal and external conditions, and result in severe decline in the facial recognition rate. To tackle this problem, a skin heat transfer (SHT) model based on thermal physiology is derived in this paper. The proposed model converts the facial thermograms into blood-perfusion data, which is revealed to reduce the within-class scatter of face images. The advantage of the derived blood-perfusion data over the raw thermograms for recognition is analyzed by the normalized reverse cumulative histogram. It is shown that blood-perfusion data are more consistent in representing facial features. The experiments conducted on both same-session and time-lapse data have further demonstrated that (1) the blood-perfusion data are less sensitive to ambient temperature, physiological and psychological conditions if the human bodies are in the steady state; (2) for time-lapse data, the performance with the blood-perfusion data is nearly identical to that of the same-session data, while the recognition rate with the temperature data dramatically decreases in this case. The major contributions of this work are the well-grounded infrared data preprocessing and the corresponding face recognition system. [Copyright &y& Elsevier]

Abstract: This paper studies the problem of determining facial orientation without correspondences in distortion-related infrared images. To improve estimation accuracy and reduce sensitivity to noise and unavoidable error, a simple and robust method based on the single linkage clustering is proposed to simultaneously detect inlier set and estimate orientation angle under contaminated data. An iterative strategy is adopted to avoid random choice of link distance in the single linkage clustering. The experimental results indicate that the proposed method is substantially superior to the moment method. This method can also be extended to detect arbitrary objects with mirror symmetrical or nearly symmetrical property. [Copyright &y& Elsevier]

In this paper, a novel learning algorithm for dynamic fuzzy neural networks based on extended radial basis function neural networks, which are functionally equivalent to Takagi–Sugeno–Kang fuzzy systems, is proposed. The algorithm comprises 4 parts: (1) criteria of rules generation; (2) allocation of premise parameters; (3) determination of consequent parameters and (4) pruning technology. The salient characteristics of the approach are: (1) a hierarchical on-line self-organizing learning paradigm is employed so that not only parameters can be adjusted, but also the determination of structure can be self-adaptive without partitioning the input space a priori; (2) fast learning speed can be achieved so that the system can be implemented in real time. Simulation studies and comprehensive comparisons with some other learning algorithms demonstrate that the proposed algorithm is superior in terms of simplicity of structure, learning efficiency and performance. [ABSTRACT FROM AUTHOR]

RRT* (Rapidly-exploring Random Tree Star), as a variant of RRT (Rapidly-exploring Random Tree), is widely used to solve path planning problems because of its asymptotic optimality. However, the algorithm is inefficient due to the high initial path cost and the slow convergence rate. In this paper, we propose a More Quickly-RRT* (MQ-RRT*) path planning algorithm based on optimized sampling points to solve the problems. A sparse sampling mechanism is proposed in MQ-RRT* to improve the global search efficiency by reducing repetitive sampling. To make the random tree oriented when expanding, a dynamic goal-biased strategy is proposed, which can reduce the sampling time. Like Q-RRT* (Quick-RRT*), MQ-RRT* expands the set of possible parent nodes in the ChooseParent and Rewire phases, which reduces the path cost. On this basis, a method for creating a new parent node close to the obstacle is proposed. The creation process can be divided into two steps: Remove-tips and CreateNodes , which further reduces the cost of path generation and makes the path smoother by using the triangle inequality principle. Finally, numerical simulations are used to compare the proposed algorithm with RRT*, Q-RRT*, GuILD (Guided Incremental Local Densification), and F-RRT* (Fast-RRT*), which verifies that the proposed algorithm has certain advantages in path cost, convergence rate, and path smoothness. [ABSTRACT FROM AUTHOR]

Audio–Visual Segmentation (AVS) aims to extract the sounding objects from a video. The current learning-based AVS methods are often supervised, which rely on specific task data annotations and expensive model training. Recognizing that the video background captured by a static camera is represented as a low-rank matrix, we introduce the non-convex robust principal component analysis into AVS task in this paper. This approach is unsupervised and only relies on input data patterns. Specifically, the proposed method decomposes each modality into the sum of two parts: the low-rank part that represents the background audio and visual information, and the sparse part that represents the foreground information. Furthermore, CUR decomposition is employed at each iteration to reduce the computational complexity in optimization. The experimental results also show that the proposed AVS outperforms the supervised methods on AVS-Bench Single-Source datasets. • Firstly introduces non-convex RPCA for Audio–Visual Segmentation. • Applies CUR decomposition to reduce computational complexity. • The task of unsupervised Audio–Visual Segmentation. • Demonstrates efficiency and effectiveness results on the AVS dataset. [ABSTRACT FROM AUTHOR]

• We proposed a novel AdaptiveWeighted Guided Image Filtering algorithm which outperforms the existing guided filters to preserve accurate edges. • We designed four special cases of the adaptive adjustment algorithm to remove the noise while enhancing depth details. • We proposed a new strategy based on AWGIF algorithm for depth enhancement in Shape-From-Focus. • Our proposed method are applied to the emerging AR/VR on hand-held devices in real-time. Existing shape from focus (SFF) techniques cannot preserve depth edges and fine structural details from a sequence of multi-focus images. Moreover, noise in the sequence affects the accuracy of the depth map. In this paper, a novel depth enhancement algorithm for the SFF based on an adaptive weighted guided image filtering (AWGIF) is proposed to address the above issues. The AWGIF is applied to decompose an initial depth map estimated by the traditional SFF into base and detail layers. In order to preserve the edges accurately in the refined depth map, the guidance image is constructed from the sequence, and the coefficient of the AWGIF is utilized to suppress the noise while enhancing the fine depth details. Experiments on real and synthetic objects demonstrate the superiority of our algorithm in terms of anti-noise, and the ability to preserve depth edges and fine structural details w.r.t. existing methods. [ABSTRACT FROM AUTHOR]

Inspired by the observation that rougher surfaces yield blurrier reflective images, a novel roughness estimation method via line blur functions (LBF) of edges reflected from such surfaces is proposed in this paper. We first reveal that the point blur function (PBF) is determined by the surface roughness base on the point blur mechanism, and mathematically analyze that the LBF can be derived from its correspondent PBF in the perpendicular direction of edges. Then, the variance of Gaussian curve which fits the LBF is mapped to the surface roughness measured by the stylus profilometry. This mapped linear function is finally applied to measure the roughness of test samples. The experimental results show that when roughness is under 0.29 μ m, the proposed method is comparable to or even higher than the stylus method in accuracy, and has advantages of convenience, high efficiency and non-contact. • Roughness of surfaces with weak textures is estimated from reflective images. • The PBF and the LBF are defined to evaluate the blur degree of reflective images. • The LBF can be derived from the correspondent PBF. • The surface roughness is formulated in terms of the LBF. [ABSTRACT FROM AUTHOR]