Fully complex-valued ELM classifiers for human action recognition

In this paper, we present a fast learning neural network classifier for human action recognition. The proposed classifier is a fully complex-valued neural network with a single hidden layer. The neurons in the hidden layer employ the fully complex-valued hyperbolic secant as an activation function. The parameters of the hidden layer are chosen randomly and the output weights are estimated analytically as a minimum norm least square solution to a set of linear equations. The fast leaning fully complex-valued neural classifier is used for recognizing human actions accurately. Optical flow-based features extracted from the video sequences are utilized to recognize 10 different human actions. The feature vectors are computationally simple first order statistics of the optical flow vectors, obtained from coarse to fine rectangular patches centered around the object. The results indicate the superior performance of the complex-valued neural classifier for action recognition. The superior performance of the complex neural network for action recognition stems from the fact that motion, by nature, consists of two components, one along each of the axes.

Efficient topological localization using orientation adjacency coherence histograms

This paper describes an efficient vision-based global topological localization approach that uses a coarse-to-fine strategy. Orientation Adjacency Coherence Histogram (OACH), a novel image feature, is proposed to improve the coarse localization. The coarse localization results are taken as inputs for the fine localization which is carried out by matching Harris-Laplace interest points characterized by the SIFT descriptor. Computation of OACHs and interest points is efficient due to the fact that these features are computed in an integrated process. We have implemented and tested the localization system in real environments. The experimental results demonstrate that our approach is efficient and reliable in both indoor and outdoor environments. © 2006 IEEE.

Vision-based global localization using a visual vocabular

This paper presents a novel coarse-to-fine global localization approach that is inspired by object recognition and text retrieval techniques. Harris-Laplace interest points characterized by SIFT descriptors are used as natural land-marks. These descriptors are indexed into two databases: an inverted index and a location database. The inverted index is built based on a visual vocabulary learned from the feature descriptors. In the location database, each location is directly represented by a set of scale invariant descriptors. The localization process consists of two stages: coarse localization and fine localization. Coarse localization from the inverted index is fast but not accurate enough; whereas localization from the location database using voting algorithm is relatively slow but more accurate. The combination of coarse and fine stages makes fast and reliable localization possible. In addition, if necessary, the localization result can be verified by epipolar geometry between the representative view in database and the view to be localized. Experimental results show that our approach is efficient and reliable. ©2005 IEEE.

Video segmentation with superpixels

Due to its importance, video segmentation has regained interest recently. However, there is no common agreement about the necessary ingredients for best performance. This work contributes a thorough analysis of various within- and between-frame affinities suitable for video segmentation. Our results show that a frame-based superpixel segmentation combined with a few motion and appearance-based affinities are sufficient to obtain good video segmentation performance. A second contribution of the paper is the extension of [1] to include motion-cues, which makes the algorithm globally aware of motion, thus improving its performance for video sequences. Finally, we contribute an extension of an established image segmentation benchmark [1] to videos, allowing coarse-to-fine video segmentations and multiple human annotations. Our results are tested on BMDS [2], and compared to existing methods. © 2013 Springer-Verlag.

基于天际线识别的无人机实时姿态角估计

以无人机天际线识别为背景,提出了一种准确、实时的天际线识别算法,并由此估计姿态角。在结合实际情况的基础上,对天际线建立能量泛函模型,利用变分原理推出相应偏微分方程。在实际应用中出于对实时性的考虑,引入直线约束对该模型进行简化,然后利用由粗到精的思想识别天际线。首先,对图像预处理并垂直剖分,然后利用简化的水平直线模型对天际线进行粗识别,通过拟合获得天际线粗识别结果,最后在基于梯度和区域混合开曲线模型约束下精确识别天际线,并由此估计无人机滚动和俯仰姿态角。实验结果表明,该算法对天际线识别具有较好的鲁棒性、准确性和实时性。

河流成因储层沉积微相与剩余油分布及形成机理研究

Based on the study of fluvial sandstone reservoir in upper of Guantao group in Gudao and Gudong oilfields, this paper first introduces A.D.Miall's(1996a) architectural-element analysis method that was summarized from ground outcrop scale into the reservoir formation research of the study area, more subtly divides sedimentary microfacies and establishes sedimentary model of research area.on this base, this paper summarizes the laws of residual oil distribution of fluvial formation and the control effect of sedimentary microfacies to residual oil distribution, and reveals residual oil formation mechanism. These results have been applied to residual oil production, and the economic effect is good. This paper will be useful for residual oil research and production and enhancement of oil recovery in similar reservoir. The major conclusions of this paper are as follows. 1. Using the architectural-element analysis method to the core data, a interfacial division scheme of the first to the dixth scale is established for the studied fluvial formation. 2.Seven architectural-elements are divided in upper of Guantao group of study area. The sandstone group 5～1+2 of Neogene upper Gutao group belongs to high sinuous fine grain meandering river, and the sandstone group 6 is sandy braided river. 3. Inter layer, the residual oil saturation of "non-main layer" is higher than "main layer", but the residual recoverable reserve of former is larger. Therefore, "main layer" is the main body of residual oil distribution. The upper and middle part of inner layer has lower permeability and strong seeping resistance. Addition to gravity effect in process of driving, its driving efficiency is low; residual oil saturation is high. Because of controlling of inside non-permeable interlayer or sedimentary construction, the residual oil saturation of non-driving or lower driving efficiency position also is high. On plane, the position of high residual oil saturation mostly is at element LV, CS, CH (FF), FF etc, Which has lower porosity and permeability, as well as lens sand-body and sand-body edge that is not controlled by well-net, non-perfect area of injection and production, lower press difference resort area of inter-well diffiuent-line and shelter from fault, local high position of small structure. 4.Microscopic residual oil mainly includes the non-moved oil in the structure of fine pore network, oil in fine pore and path, oil segment in pore and path vertical to flow direction, oil spot or oil film in big pore, residual oil in non-connective pore. 5.The most essential and internal controlling factor of fluvial formation residual oil distribution is sedimentary microfacies. Status of injection and production is the exterior controlling factor of residual oil distribution. 6. The controlling effect of formation sedimentary microfacies to residual oil distribution indicates inter-layer vertical sedimentary facies change in scale of injection and production layer-series, planar sedimentary face change and inner-layer vertical sedimentary rhythm and interbed in single layer to residual oil distribution. 7. It is difficult to clear up the inter-layer difference in scale of injection and production layer-series. The using status of minor layer is not good and its residual oil saturation is high relatively. It is obvious that inter-layer vertical sedimentary facies changes control inter-layer residual oil distribution at the same or similar conditions of injection and production. For fluvial formation, this vertical sedimentary facies change mainly is positive gyration. Namely, from down to top, channel sediment (element CHL, LA) changes into over-bank sediment (element LV, CR, CS). 8. In water-injection developing process of transverse connecting fluvial sandstone oil formation, injection water always comes into channel nearby, and breaks through along channel and orientation of high pressure gradient, does not expand into side of channel until pressure gradient of channel orientation changes into low. It brings about that water-driving status of over-bank sedimentary element formation (LV, CR, CS) is not good, residual oil saturation is high. In non-connective abandoned channel element (CH) formation with channel, because this reverse is difficult to control by injection and production well-series, its using status is not good, even terribly not good, residual oil is enrichment. 9. The rhythm and sedimentary structure, sedimentary facies change in single sand body brings about vertical changes of formation character, growth character of inner layer interbed. These are important factor of controlling and affecting vertical water spread volume and inner layer residual oil forming and distribution in single sand body. Positive rhythm, is the principal part of fluvial sandstone inner layer sedimentary rhythm. Namely, from down to upside, rock grain granularity changes from coarse to fine, seeping ability changes from strong to feebleness. It brings about that water-driving status of inner layer upside is not good, residual oil saturation is high. Inner layer interbed has different degree affecting and controlling effect to seeping of oil and water. Its affecting degree lies on interbed thickness, extending scale, position, and jeted segment of production or injection well. The effect of interbed at upside of oil formation to oil and water seeping is less; the effect of interbed at middle of oil formation to oil and water seeping is more. 10. Indoor experiment and research indicate that wettability, permeability step, vertical permeability, position of Kmax and ratio of oil viscousity and water viscousity all have great effect on the water-driving recovery ratio. 11. Microscopic residual oil distribution is affected and controlled by formation pore network structure, pressure field distribution, and oil characteristic. 12.The residual oil forming mechanism: the over-bank sedimentary element and upper part of a positive rhythm sandstone have fine pore and throat network, permeability is low, displacement pressure of pore and throat is high. The water-driving power usually falls short of displacement pressure that brings about injection water does not spread into these pore and throat network, thereby immovable oil area, namely residual oil, is formed. At underside of channel sedimentary element and positive rhythm sandstone, porosity and permeability is relatively high, connecting degree of pore and throat is high, displacement pressure of pore and throat is low. Thereby injection water is easy to enter into pore and throat, driving oil in them. Because the pore space is irregular, the surface of pore wall is coarse and non-flat. That the oil locate on concave hole of pore wall and the dead angle of pore, and the oil attaches on surface of pore wall by surface tension, are difficult to be peeled off, becoming water-driving residual oil (remaining oil). On the other hand, Because flowing section lessens, flowing resistance increase, action of capillary fore, or seeping speed decreases at process of transfer at pass narrow throat path in the course carried by driving water. The "oil drop", "oil bead", or "oil segment" peeled off by driving water is difficult to carry and to drive out by water at less pressure difference. Thereby they are enclosed in pore to form discontinuous residual oil. 13.This results described above have been applied in nine develop blocks of Gudao and Gudong oilfield. Its applying effect is marked through local injection production adjustment, deploying replacement well, repair hole, replacement envelop, block off water and profile control etc. Relative method and technology can be applied to other oil production area of Shengli oilfield, and obtain better economic and societal effect.

Vectorizing Face Images by Interpreting Shape and Texture Computations

The correspondence problem in computer vision is basically a matching task between two or more sets of features. In this paper, we introduce a vectorized image representation, which is a feature-based representation where correspondence has been established with respect to a reference image. This representation has two components: (1) shape, or (x, y) feature locations, and (2) texture, defined as the image grey levels mapped onto the standard reference image. This paper explores an automatic technique for "vectorizing" face images. Our face vectorizer alternates back and forth between computation steps for shape and texture, and a key idea is to structure the two computations so that each one uses the output of the other. A hierarchical coarse-to-fine implementation is discussed, and applications are presented to the problems of facial feature detection and registration of two arbitrary faces.

Index Trees for Efficient Deformable Shape-based Retrieval

An improved method for deformable shape-based image indexing and retrieval is described. A pre-computed index tree is used to improve the speed of our previously reported on-line model fitting method; simple shape features are used as keys in a pre-generated index tree of model instances. In addition, a coarse to fine indexing scheme is used at different levels of the tree to further improve speed while maintaining matching accuracy. Experimental results show that the speedup is significant, while accuracy of shape-based indexing is maintained. A method for shape population-based retrieval is also described. The method allows query formulation based on the population distributions of shapes in each image. Results of population-based image queries for a database of blood cell micrographs are shown.

Interactive quadrangulation with reeb atlases and connectivity textures

Creating high-quality quad meshes from triangulated surfaces is a highly nontrivial task that necessitates consideration of various application specific metrics of quality. In our work, we follow the premise that automatic reconstruction techniques may not generate outputs meeting all the subjective quality expectations of the user. Instead, we put the user at the center of the process by providing a flexible, interactive approach to quadrangulation design. By combining scalar field topology and combinatorial connectivity techniques, we present a new framework, following a coarse to fine design philosophy, which allows for explicit control of the subjective quality criteria on the output quad mesh, at interactive rates. Our quadrangulation framework uses the new notion of Reeb atlas editing, to define with a small amount of interactions a coarse quadrangulation of the model, capturing the main features of the shape, with user prescribed extraordinary vertices and alignment. Fine grain tuning is easily achieved with the notion of connectivity texturing, which allows for additional extraordinary vertices specification and explicit feature alignment, to capture the high-frequency geometries. Experiments demonstrate the interactivity and flexibility of our approach, as well as its ability to generate quad meshes of arbitrary resolution with high-quality statistics, while meeting the user's own subjective requirements.