PMSC: PatchMatch-Based Superpixel Cut for Accurate Stereo Matching

Do we really need 3D labels in order to learn how to predict 3D? In this paper, we show that one can learn a mapping from appearance to 3D properties without ever seeing a single explicit 3D label. Rather than use explicit supervision, we use the regularity of indoor scenes to learn the mapping in a completely unsupervised manner. We demonstrate this on both a standard 3D scene understanding dataset as well as Internet images for which 3D is unavailable, precluding supervised learning. Despite never seeing a 3D label, our method produces competitive results.

Outlier detection and disparity refinement in stereo matching

The problem addressed here is the challenge of determining the best production strategy for multilayer reservoirs which varies from field to field as well as from well to well. Commingled production strategy is one of the most commonly used for several multilayer wells. However, there are some limitations to this method. Mostly, when very complex multilayer wells which contain different types of fluid in each productive layer are identified, the selection process of production approach is basically trial and error and later found to be a poor choice. One attempted solution to choosing the optimal production strategy is using numerical reservoir simulation. In this study, we used reservoir simulation to compare the performance of four different production strategies for different combinations of types of fluid present in a two-layer system. Strategy 1 in which the fluid is produced from two layers simultaneously is the best production approach when oil is present in the top and bottom layers. Strategy 2 in which the fluid is produced from one layer at a time starting with the bottom layer is the best production approach when reservoir fluid at the top layer is gas while fluids present in the bottom layer can vary from oil, gas&oil, gas&water and oil&water. Strategy 3 produces fluid from one layer at a time, starting with the bottom layer. When the production rate decreases to half of the original rate, the two zones are produced in a commingled fashion. Similar to strategy 1, strategy 3 is also the best production approach when oil is present in the top and bottom layers. Strategy 4 is similar to strategy 3 but starting with the top layer. This strategy is the best production approach when oil is present in the top layer and gas&water in the bottom layer. This approach can delay water production and sustain reservoir pressure.

This paper presents a stereo matching system using a codebook for natural user interface (NUI). Both color and disparity information in previous frames are stored and compared to deal with the occlusion problem in stereo matching.

The stereo matching <span>process is one of the key areas that impact the stereo vision technologies which are commonly used in the application of three-dimensional reconstructions. The accuracy of the depth information used in three-dimensional reconstruction is directly proportional to the accuracy of the disparity obtained from stereo matching. The challenging issue in the stereo matching process is to determine the accurate corresponding point between the left image and right image, especially for image pairs that have different exposure such as different illumination and image pair with less texture region. In order to increase the accuracy of disparity value, a new stereo matching algorithm is proposed based on the combination of Sum of absolute different and census transform at matching cost computation. guided filter was used in the matching cost aggregation in order to remove noise and preserve the edge of the image. In the optimization step, the winner take all strategy is used to select the minimum matching cost. Finally, a median filter is applied to the initial disparity map for refinement purposes. The experimental results show that the algorithm is effective in reducing the error and improving the accuracy of the disparity map in different illumination regions, less textured regions and different environmental exposure.</span>

Wavelength discrimination (WLD) detector optimization for time-of-flight positron emission tomography with depth of interaction information

This paper reports the results from numerical modeling of spudcan deep penetration through three-layer clays. Two typical seabed strength profiles commonly encountered in the field, and identified as critical for potential punch-through failure, were considered: (1) uniform stiff-soft-stiff clay and (2) nonuniform clay with an interbedded stiff clay layer. Three-dimensional large deformation finite-element (LDFE) analyses were carried out with and without simulating strain softening and strain rate dependency of the shear strength. The results were compared with previously published LDFE results and centrifuge test data. A detailed parametric study was undertaken, varying the relevant range of layer thicknesses (relative to the spudcan diameter), strength ratios, normalized strength, and strength nonhomogeneity. Punch-through and rapid leg penetration (for stiff-over-soft) and squeezing (for the reverse) were demonstrated by the penetration resistance profiles and associated soil failure mechanisms. The results emphasized that the resistance profile from a spudcan continuous penetration in a three-layer sediment (e.g., stiff-soft-stiff) may be significantly different from the combined profiles resulting from two separate analyses on first-second layers and second-third layers, even if an appropriate depth factor is accounted for. For uniform stiff-soft-stiff clays, the presence of the bottom stiff clay layer within a short distance of one diameter affected spudcan bearing response in the top layer and suppressed the potential for punch-through failure (apart from rapid leg run for t1/D=0.75 and su1s/su2=3). A trapped soil plug beneath the advancing spudcan, following penetration of the top stiff layer, caused the bottom stiff layer to be sensed earlier and led to enhancement of the limiting squeezing depth. For nonuniform clay with an interbedded stiff layer, spudcan bearing response was dominated by squeezing (with limiting squeezing depth of approximately 0.18D) in the top soft layer and potential punch-through and rapid leg run (for t2/D≥0.5 and su2/su3s=2.5) in the middle and bottom layers. In contrast to ISO recommendations, the soft soils in between the advancing spudcan base and the stronger layer did not squeeze out completely. Instead, some of the trapped material was forced into the underlying stronger layer. Punch-through distance was found to reduce significantly with increasing positive strength gradient of the bottom layer.

Partial label learning deals with the problem where each training example is associated with a set of candidate labels, only one of which is assumed to be valid. To learn from such ambiguous labeling information, the critical point is to disambiguate the set of candidate labels, thereby targeting the ground-truth label. By utilizing the nature that only one of the candidate labels is correct, we employ the entropy minimization strategy to force the model making confident predictions of the training data. By doing this, the ground-truth labels are likely to make more contributions to the model training. Finally, comparative experiments on a number of real-world datasets are conducted, demonstrating the effectiveness of the proposed approach.

The Euclidean embedding learning based on convolutional neural network for stereo matching

Although convolutional neural network (CNN)-based stereo matching methods have become increasingly popular thanks to their robustness, they primarily have been focused on the matching cost computation. By leveraging CNNs, we present a novel method for matching cost aggregation to boost the stereo matching performance. Our insight is to learn the convolution kernel within CNN architecture for cost aggregation in a fully convolutional manner. Tailored to cost aggregation problem, our method differs from handcrafted methods in terms of its convolutional aggregation through optimally learned CNNs. First, the matching cost is aggregated with cost volume unary network, and then optimized with explicit disparity boundary, estimated through disparity boundary pairwise network, within a global energy minimization. Experiments demonstrate that our method outperforms conventional hand-crafted aggregation methods.

Partial multi-label learning (PML) models the scenario where each training instance is annotated with a set of candidate labels, and only some of the labels are relevant. The PML problem is practical in real-world scenarios, as it is difficult and even impossible to obtain precisely labeled samples. Several PML solutions have been proposed to combat with the prone misled by the irrelevant labels concealed in the candidate labels, but they generally focus on the smoothness assumption in feature space or low-rank assumption in label space, while ignore the negative information between features and labels. Specifically, if two instances have largely overlapped candidate labels, irrespective of their feature similarity, their ground-truth labels should be similar; while if they are dissimilar in the feature and candidate label space, their ground-truth labels should be dissimilar with each other. To achieve a credible predictor on PML data, we propose a novel approach called PML-LFC (Partial Multi-label Learning with Label and Feature Collaboration). PML-LFC estimates the confidence values of relevant labels for each instance using the similarity from both the label and feature spaces, and trains the desired predictor with the estimated confidence values. PML-LFC achieves the predictor and the latent label matrix in a reciprocal reinforce manner by a unified model, and develops an alternative optimization procedure to optimize them. Extensive empirical study on both synthetic and real-world datasets demonstrates the superiority of PML-LFC.

Thermo-physiological comfort of the fabric is achieved through the ability to manage heat and transmit sensible and insensible perspiration through the fabric. Ergonomic comfort aspects of sports clothing can be improved by enhancing the moisture management properties of knitted fabrics. Double-face knitted structures were developed by selecting the right type of material for the top and bottom layer. In this paper, the moisture management properties of four double-face knitted fabrics with hydrophobic fibre in the top layer (Polypropylene and micro-fibre polyester) and hydrophilic fibre in the bottom layer (Bamboo) were studied to find out the suitability for sportswear. The subjective perception was carried out using thermal state parameters and was correlated with Overall Moisture Management Capacity (OMMC). Double-face fabric knitted with micro-fibre polyester in the top layer and bamboo in the bottom layer has excellent moisture management properties which seem suitable for next-to-skin application. However, micro-fibre polyester (20%)/polypropylene (20%) yarns in the top layer on account of polypropylene due to its dry feel shows good OMMC followed by micro-fibre polyester in the top layer. Univariate analysis of variance with a confidence level of 95% showed that all the moisture management indices are statistically significant.

Current-induced switching in CoGa/L10 MnGa/(CoGa)/Pt structure with different thicknesses

The paper focuses on estimating disparity map that can handle outliers. Outliers are the major challenging problem in stereo matching. The results obtained from depth maps are significantly affected by outliers regions that are present in the image. The stereo images are pre-processed using homomorphic filter. The Semi-global block matching method is used to estimate disparity information. Left to right consistency check method is applied to identify outliers. Outliers identified are classified as occlusion and mismatches and then they are refined. The proposed method is robust against illumination differences and recovers disparities at the object edges, occluded regions. The performance of the technique used is analyzed by percentage of bad matching pixel and root mean square error. Experimental results generated determine that the proposed algorithm gives more stable disparity maps.

Traditionally, the problem of stereo matching has been addressed either by a sparse feature-based approach or a dense area-based approach. This technique still suffers from a lack in accuracy or long computation time needed to match stereo images. In this paper, we employ color segments on scan lines of stereo images as primitives for matching, and introduce the widely used fuzzy set theory to the matching processes in order to handle ambiguities in matching. During the stage of accurate pixel correspondences of matched color segments, we use the technique of dynamic time warping (DTW), which is used in speech recognition. We found that DTW works robustly and rapidly regardless of the number of points to be matched. Covering more image space than point elements and providing more information than gray level features, color segments are less sensitive to noise and photometric variations and are thus more easily matched, and can reduce the false matches. Furthermore, a post-processing step can find more matches and alleviate the influence of under-segmentation or over-segmentation. A fast median filter can be applied to the results to further remove outliers. Experiments have been performed to demonstrate the accuracy and efficiency of the algorithm.