“Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks”の学習報告

Travel time is a fundamental measure in transportation. Accurate travel-time prediction also is crucial to the development of intelligent transportation systems and advanced traveler information systems. We apply support vector regression (SVR) for travel-time prediction and compare its results to other baseline travel-time prediction methods using real highway traffic data. Since support vector machines have greater generalization ability and guarantee global minima for given training data, it is believed that SVR will perform well for time series analysis. Compared to other baseline predictors, our results show that the SVR predictor can significantly reduce both relative mean errors and root-mean-squared errors of predicted travel times. We demonstrate the feasibility of applying SVR in travel-time prediction and prove that SVR is applicable and performs well for traffic data analysis.

/pdf/travel-time-prediction-with-support-vector-regression-3nxajn3hpe.pdf

Travel-time prediction with support vector regression

Automatic license plate recognition (LPR) plays an important role in numerous applications and a number of techniques have been proposed. However, most of them worked under restricted conditions, such as fixed illumination, limited vehicle speed, designated routes, and stationary backgrounds. In this study, as few constraints as possible on the working environment are considered. The proposed LPR technique consists of two main modules: a license plate locating module and a license number identification module. The former characterized by fuzzy disciplines attempts to extract license plates from an input image, while the latter conceptualized in terms of neural subjects aims to identify the number present in a license plate. Experiments have been conducted for the respective modules. In the experiment on locating license plates, 1088 images taken from various scenes and under different conditions were employed. Of which, 23 images have been failed to locate the license plates present in the images; the license plate location rate of success is 97.9%. In the experiment on identifying license number, 1065 images, from which license plates have been successfully located, were used. Of which, 47 images have been failed to identify the numbers of the license plates located in the images; the identification rate of success is 95.6%. Combining the above two rates, the overall rate of success for our LPR algorithm is 93.7%.

/pdf/automatic-license-plate-recognition-c8wfmp9nt1.pdf

Automatic license plate recognition

Uncertainty quantification (UQ) plays a pivotal role in reduction of uncertainties during both optimization and decision making processes. It can be applied to solve a variety of real-world applications in science and engineering. Bayesian approximation and ensemble learning techniques are two most widely-used UQ methods in the literature. In this regard, researchers have proposed different UQ methods and examined their performance in a variety of applications such as computer vision (e.g., self-driving cars and object detection), image processing (e.g., image restoration), medical image analysis (e.g., medical image classification and segmentation), natural language processing (e.g., text classification, social media texts and recidivism risk-scoring), bioinformatics, etc. This study reviews recent advances in UQ methods used in deep learning. Moreover, we also investigate the application of these methods in reinforcement learning (RL). Then, we outline a few important applications of UQ methods. Finally, we briefly highlight the fundamental research challenges faced by UQ methods and discuss the future research directions in this field.

A Review of Uncertainty Quantification in Deep Learning: Techniques, Applications and Challenges

Traditional and recent approaches in background modeling for foreground detection: An overview

Structures matter in single image super resolution (SISR). Recent studies benefiting from generative adversarial network (GAN) have promoted the development of SISR by recovering photo-realistic images. However, there are always undesired structural distortions in the recovered images. In this paper, we propose a structure-preserving super resolution method to alleviate the above issue while maintaining the merits of GAN-based methods to generate perceptual-pleasant details. Specifically, we exploit gradient maps of images to guide the recovery in two aspects. On the one hand, we restore high-resolution gradient maps by a gradient branch to provide additional structure priors for the SR process. On the other hand, we propose a gradient loss which imposes a second-order restriction on the super-resolved images. Along with the previous image-space loss functions, the gradient-space objectives help generative networks concentrate more on geometric structures. Moreover, our method is model-agnostic, which can be potentially used for off-the-shelf SR networks. Experimental results show that we achieve the best PI and LPIPS performance and meanwhile comparable PSNR and SSIM compared with state-of-the-art perceptual-driven SR methods. Visual results demonstrate our superiority in restoring structures while generating natural SR images.

/pdf/structure-preserving-super-resolution-with-gradient-guidance-3jqztf94h3.pdf

Structure-Preserving Super Resolution With Gradient Guidance

The precise localization of 3D objects from a single image without depth information is a highly challenging problem. Most existing methods adopt the same approach for all objects regardless of their diverse distributions, leading to limited performance for truncated objects. In this paper, we propose a flexible framework for monocular 3D object detection which explicitly decouples the truncated objects and adaptively combines multiple approaches for object depth estimation. Specifically, we decouple the edge of the feature map for predicting long-tail truncated objects so that the optimization of normal objects is not influenced. Furthermore, we formulate the object depth estimation as an uncertainty-guided ensemble of directly regressed object depth and solved depths from different groups of keypoints. Experiments demonstrate that our method outperforms the state-of-the-art method by relatively 27% for the moderate level and 30% for the hard level in the test set of KITTI benchmark while maintaining real-time efficiency. Code will be available at https://github.com/zhangyp15/MonoFlex.

/pdf/objects-are-different-flexible-monocular-3d-object-detection-32p93ibhpl.pdf

Objects are Different: Flexible Monocular 3D Object Detection

Structure-Preserving Super Resolution with Gradient Guidance

In this paper, we propose to learn a deep fitting degree scoring network for monocular 3D object detection, which aims to score fitting degree between proposals and object conclusively. Different from most existing monocular frameworks which use tight constraint to get 3D location, our approach achieves high-precision localization through measuring the visual fitting degree between the projected 3D proposals and the object. We first regress the dimension and orientation of the object using an anchor-based method so that a suitable 3D proposal can be constructed. We propose FQNet, which can infer the 3D IoU between the 3D proposals and the object solely based on 2D cues. Therefore, during the detection process, we sample a large number of candidates in the 3D space and project these 3D bounding boxes on 2D image individually. The best candidate can be picked out by simply exploring the spatial overlap between proposals and the object, in the form of the output 3D IoU score of FQNet. Experiments on the KITTI dataset demonstrate the effectiveness of our framework.

/pdf/deep-fitting-degree-scoring-network-for-monocular-3d-object-raun30a3kk.pdf

Deep Fitting Degree Scoring Network for Monocular 3D Object Detection

In this paper, we contribute a new million-scale face benchmark containing noisy 4M identities/260M faces (WebFace260M) and cleaned 2M identities/42M faces (WebFace42M) training data, as well as an elaborately designed time-constrained evaluation protocol. Firstly, we collect 4M name list and download 260M faces from the Internet. Then, a Cleaning Automatically utilizing Self-Training (CAST) pipeline is devised to purify the tremendous WebFace260M, which is efficient and scalable. To the best of our knowledge, the cleaned WebFace42M is the largest public face recognition training set and we expect to close the data gap between academia and industry. Referring to practical scenarios, Face Recognition Under Inference Time conStraint (FRUITS) protocol and a test set are constructed to comprehensively evaluate face matchers.Equipped with this benchmark, we delve into million-scale face recognition problems. A distributed framework is developed to train face recognition models efficiently without tampering with the performance. Empowered by Web-Face42M, we reduce relative 40% failure rate on the challenging IJB-C set, and rank the 3rd among 430 entries on NIST-FRVT. Even 10% data (WebFace4M) shows superior performance compared with public training set. Furthermore, comprehensive baselines are established on our rich-attribute test set under FRUITS-100ms/500ms/1000ms protocol, including MobileNet, EfficientNet, AttentionNet, ResNet, SENet, ResNeXt and RegNet families. Benchmark website is https://www.face-benchmark.org.

/pdf/webface260m-a-benchmark-unveiling-the-power-of-million-scale-1wx2pui5qe.pdf

Jie Zhou

Papers

Structure-Preserving Super Resolution With Gradient Guidance

Objects are Different: Flexible Monocular 3D Object Detection

Structure-Preserving Super Resolution with Gradient Guidance

Deep Fitting Degree Scoring Network for Monocular 3D Object Detection

WebFace260M: A Benchmark Unveiling the Power of Million-Scale Deep Face Recognition