Object localization in 3D space is a challenging aspect in monocular 3D object detection. Recent advances in 6DoF pose estimation have shown that predicting dense 2D-3D correspondence maps between image and object 3D model and then estimating object pose via Perspective-n-Point (PnP) algorithm can achieve remarkable localization accuracy. Yet these methods rely on training with ground truth of object geometry, which is difficult to acquire in real outdoor scenes. To address this issue, we propose MonoRUn, a novel detection framework that learns dense correspondences and geometry in a self-supervised manner, with simple 3D bounding box annotations. To regress the pixel-related 3D object coordinates, we employ a regional reconstruction network with uncertainty awareness. For self-supervised training, the predicted 3D coordinates are projected back to the image plane. A Robust KL loss is proposed to minimize the uncertainty-weighted reprojection error. During testing phase, we exploit the network uncertainty by propagating it through all downstream modules. More specifically, the uncertainty-driven PnP algorithm is leveraged to estimate object pose and its covariance. Extensive experiments demonstrate that our proposed approach outperforms current state-of-the-art methods on KITTI benchmark.1

For vehicles equipped with the automatic parking system, the accuracy and speed of the parking slot detection are crucial. But the high accuracy is obtained at the price of low speed or expensive computation equipment, which are sensitive for many car manufacturers. In this paper, we proposed a detector using CNN (convolutional neural networks) for faster speed and smaller model size while keeps accuracy. To achieve the optimal balance, we developed a strategy to select the best receptive fields and prune the redundant channels automatically after each training epoch. The proposed model is capable of jointly detecting corners and line features of parking slots while running efficiently in real time on average processors. The model has a frame rate of about 30 FPS on a 2.3 GHz CPU core, yielding parking slot corner localization error of 1.51±2.14 cm (std. err.) and slot detection accuracy of 98%, generally satisfying the requirements in both speed and accuracy on onboard mobile terminals.

MOTIVATION: As the use of microarrays in human studies continues to increase, stringent quality assurance is necessary to ensure accurate experimental interpretation. We present a formal approach for microarray quality assessment that is based on dimension reduction of established measures of signal and noise components of expression followed by parametric multivariate outlier testing. RESULTS: We applied our approach to several data resources. First, as a negative control, we found that the Affymetrix and Illumina contributions to MAQC data were free from outliers at a nominal outlier flagging rate of alpha=0.01. Second, we created a tunable framework for artificially corrupting intensity data from the Affymetrix Latin Square spike-in experiment to allow investigation of sensitivity and specificity of quality assurance (QA) criteria. Third, we applied the procedure to 507 Affymetrix microarray GeneChips processed with RNA from human peripheral blood samples. We show that exclusion of arrays by this approach substantially increases inferential power, or the ability to detect differential expression, in large clinical studies. AVAILABILITY: http://bioconductor.org/packages/2.3/bioc/html/arrayMvout.html and http://bioconductor.org/packages/2.3/bioc/html/affyContam.html affyContam (credentials: readonly/readonly)

As the Web expands exponentially, there are a flood of pornographic Web sites on the Internet. Thus effective and fast web filtering systems are essential. Web filtering based on hypertext classification has become one of the important techniques to handle and filter inappropriate information on the Web. The task involved can be parallelized and distributed in a grid environment. However, how to improve the performance of the hypertext classification under the situation of noisy data is still a challenging problem. In this paper, we propose a new approach for hypertext classification in Web filtering, which uses a novel support vector machine and k-nearest neighbor (KNN-SVM) to remove noisy training examples. The task of text categorization is distributed in several computers. The experimental results show that the generalization performance in the accuracy of classification and the processing time are improved significantly compared to that of the traditional SVM classifier over the grid, and adapt to engineering applications.

Port state control (PSC) inspection is the most important mechanism to ensure world marine safe. Recently, some SVM-based risk assessment systems have been presented in the world. They estimate the risk of each candidate ship based on its generic factors and history inspection factors to select high-risk one before conducting on-board PSC inspection. However, how to improve the performance of the PSC inspection under the situation of noisy data when applying SVM is still a challenging problem. In this paper, we propose a new approach for PSC inspection, which uses a novel support vector machine and k-nearest neighbor (KNN-SVM) to remove noisy training examples and Bag of Words (BW) to extract some new target factors for the PSC inspection database. The experimental results show that the generalization performance and the accuracy of risk assessment are improved significantly compared to that of the traditional SVM classifier, and adapt to engineering applications.