检索结果-内蒙古大学图书馆

Action Candidate Driven Clipped Double Q-learning for Discrete and Continuous Action Tasks

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Jiang, Haobo Xie, Jin Yang, Jian Pca Lab Key Lab of Intelligent Percept. and Syst. for High-Dimensional Information of Ministry of Education Jiangsu Key Lab of Image and Video Understanding for Social Security School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing China

Double Q-learning is a popular reinforcement learning algorithm in Markov decision process (MDP) problems. Clipped Double Q-learning, as an effective variant of Double Qlearning, employs the clipped double estimator to approximate the maximum expected action value. Due to the underestimation bias of the clipped double estimator, the performance of clipped Double Q-learning may be degraded in some stochastic environments. In this paper, in order to reduce the underestimation bias, we propose an action candidate-based clipped double estimator for Double Q-learning. Specifically, we first select a set of elite action candidates with high action values from one set of estimators. Then, among these candidates, we choose the highest valued action from the other set of estimators. Finally, we use the maximum value in the second set of estimators to clip the action value of the chosen action in the first set of estimators and the clipped value is used for approximating the maximum expected action value. Theoretically, the underestimation bias in our clipped Double Q-learning decays monotonically as the number of action candidates decreases. Moreover, the number of action candidates controls the trade-off between the overestimation and underestimation biases. In addition, we also extend our clipped Double Q-learning to continuous action tasks via approximating the elite continuous action candidates. We empirically verify that our algorithm can more accurately estimate the maximum expected action value on some toy environments and yield good performance on several benchmark problems. Code is available at https://***/Jiang-HB/AC-CDQ. Copyright © 2022, The Authors. All rights reserved.

关键词： Reinforcement learning

Bi-level Alignment for Cross-Domain Crowd Counting

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Gong, Shenjian Zhang, Shanshan Yang, Jian Dai, Dengxin Schiele, Bernt PCA Lab Key Lab of Intelligent Perception and Systems for High-Dimensional Information Ministry of Education Jiangsu Key Lab of Image and Video Understanding for Social Security School of Computer Science and Engineering Nanjing University of Science and Technology China MPI Informatics Germany

Recently, crowd density estimation has received increasing attention. The main challenge for this task is to achieve high-quality manual annotations on a large amount of training data. To avoid reliance on such annotations, previous works apply unsupervised domain adaptation (UDA) techniques by transferring knowledge learned from easily accessible synthetic data to real-world datasets. However, current state-of-the-art methods either rely on external data for training an auxiliary task or apply an expensive coarse-to-fine estimation. In this work, we aim to develop a new adversarial learning based method, which is simple and efficient to apply. To reduce the domain gap between the synthetic and real data, we design a bi-level alignment framework (BLA) consisting of (1) task-driven data alignment and (2) fine-grained feature alignment. In contrast to previous domain augmentation methods, we introduce AutoML to search for an optimal transform on source, which well serves for the downstream task. On the other hand, we do fine-grained alignment for foreground and background separately to alleviate the alignment difficulty. We evaluate our approach on five real-world crowd counting benchmarks, where we outperform existing approaches by a large margin. Also, our approach is simple, easy to implement and efficient to apply. The code is publicly available at https://***/Yankeegsj/BLA. © 2022, CC BY.

关键词： Alignment

DSFD: Dual shot face detector

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Li, Jian Wang, Yabiao Wang, Changan Tai, Ying Qian, Jianjun Yang, Jian Wang, Chengjie Li, Jilin Huang, Feiyue PCA Lab Key Lab of Intelligent Perception and Systems for High-Dimensional Information of Ministry of Education Jiangsu Key Lab of Image Video Understanding for Social Security School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing China Youtu Lab Tencent

In this paper, we propose a novel face detection network with three novel contributions that address three key aspects of face detection, including better feature learning, progressive loss design and anchor assign based data augmentation, respectively. First, we propose a Feature Enhance Module (FEM) for enhancing the original feature maps to extend the single shot detector to dual shot detector. Second, we adopt Progressive Anchor Loss (PAL) computed by two different sets of anchors to effectively facilitate the features. Third, we use an Improved Anchor Matching (IAM) by integrating novel anchor assign strategy into data augmentation to provide better initialization for the regressor. Since these techniques are all related to the two-stream design, we name the proposed network as Dual Shot Face Detector (DSFD). Extensive experiments on popular benchmarks, WIDER FACE and FDDB, demonstrate the superiority of DSFD over the state-of-the-art face detectors. Copyright © 2018, The Authors. All rights reserved.

关键词： Face recognition

Edge direction dispersion and its application in automatic suitable-matching area selection

学校读者我要写书评

暂无评论

Edge direction dispersion and its application in automatic s...

Chinese Control and Decision Conference

作者： Qing-Hai, Ding Xin-Rong, Yu Hai-Bo, Luo Hong-Mei, Liu Research Institute of General Development and Demonstration of Equipment Equipment of Air Force Bejing 100076 China Shenyang Institute of Automation Chinese Academy of Sciences Shenyang 110016 China Graduate School Chinese Academy of Sciences Beijing 100039 China Key Laboratory of Optical-Electronics Information Processing Chinese Academy of Science Shenyang 110016 China Key Laboratory of Image Understanding Computer Vision Liaoning Province 110016 China AVIC HONGDU Aviation Industry Group Ltd. Nanchang 330024 China

ISBN: (纸本)9781424487363

Automatic suitable-matching area selection is an important research in visual guidance, scene matching and local non-texture target tracking. Meanwhile, it is the major research subject in computer vision and accurate guidance. In this paper, a new characteristic quantity, i.e., the edge direction dispersion was defined. Then the computation method, parameters selection and real-time realization of this characteristic quantity were investigated. Furthermore, the application of its characteristic in automatic suitable-matching area selection was studied. With a reasonable design for calculation precision, solving method for arctan(x) function in [-∞,+∞ ] could be simplified to linearization calculations in [0, 0.25], then a fast algorithm for arctan(x) function was presented. Experimental results showed that the sub-image extracted by our fast algorithm had higher matching level and was suitable to be a matching template. All the calculations involved in the proposed matching algorithm could be done in one filed period and the real-time request could be satisfied. Moreover, this characteristic of edge direction could also be used to measure the comparability of template and real-time images, which could be widely used in automatic target recognition and target tracking. © 2011 IEEE.

关键词： Target tracking

Multi-Modal Curriculum Learning over Graphs

学校读者我要写书评

暂无评论

ACM Transactions on Intelligent Systems and Technology 2019年第4期10卷 1–25页

作者： Chen Gong Jian Yang Dacheng Tao PCA Lab Key Lab of Intelligent Perception and Systems for High-Dimensional Information of Ministry of Education and Jiangsu Key Lab of Image and Video Understanding for Social Security School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing China UBTECH Sydney Artificial Intelligence Centre and the School of Computer Science Faculty of Engineering and Information Technologies the University of Sydney Sydney Australia

Curriculum Learning (CL) is a recently proposed learning paradigm that aims to achieve satisfactory performance by properly organizing the learning sequence from simple curriculum examples to more difficult ones. Up to now, few works have been done to explore CL for the data with graph structure. Therefore, this article proposes a novel CL algorithm that can be utilized to guide the label Propagation (LP) over graphs, of which the target is to “learn” the labels of unlabeled examples on the graphs. Specifically, we assume that different unlabeled examples have different levels of difficulty for propagation, and their label learning should follow a simple-to-difficult sequence with the updated curricula. Furthermore, considering that the practical data are often characterized by multiple modalities, every modality in our method is associated with a “teacher” that not only evaluates the difficulties of examples from its own viewpoint, but also cooperates with other teachers to generate the overall simplest curriculum examples for propagation. By taking the curriculums suggested by the teachers as a whole, the common preference (i.e., commonality) of teachers on selecting the simplest examples can be discovered by a row-sparse matrix, and their distinct opinions (i.e., individuality) are captured by a sparse noise matrix. As a result, an accurate curriculum sequence can be established and the propagation quality can thus be improved. Theoretically, we prove that the propagation risk bound is closely related to the examples’ difficulty information, and empirically, we show that our method can generate higher accuracy than the state-of-the-art CL approach and LP algorithms on various multi-modal tasks.

关键词： Curriculum learning label propagation multi-modal learning semi-supervised learning

Non-aligned Supervision for Real image Dehazing

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Fan, Junkai Guo, Fei Qian, Jianjun Li, Xiang Li, Jun Yang, Jian Key Lab of Intelligent Perception and Systems for HighDimensional Information of Ministry of Education Jiangsu Key Lab of Image and Video Understanding for Social Security PCA Lab School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing China College of Computer Science Nankai University Tianjin China

Removing haze from real-world images is challenging due to unpredictable weather conditions, resulting in the misalignment of hazy and clear image pairs. In this paper, we propose an innovative dehazing framework that operates under non-aligned supervision. This framework is grounded in the atmospheric scattering model, and consists of three interconnected networks: dehazing, airlight, and transmission networks. In particular, we explore a non-alignment scenario that a clear reference image, unaligned with the input hazy image, is utilized to supervise the dehazing network. To implement this, we present a multi-scale reference loss that compares the feature representations between the referred image and the dehazed output. Our scenario makes it easier to collect hazy/clear image pairs in real-world environments, even under conditions of misalignment and shift views. To showcase the effectiveness of our scenario, we have collected a new hazy dataset including 415 image pairs captured by mobile Phone in both rural and urban areas, called"Phone-Hazy". Furthermore, we introduce a self-attention network based on mean and variance for modeling real infinite airlight, using the dark channel prior as positional guidance. Additionally, a channel attention network is employed to estimate the three-channel transmission. Experimental results demonstrate the superior performance of our framework over existing state-of-the-art techniques in the real-world image dehazing task. Phone-Hazy and code will be available at https://***/projectpage/NSDNet/***. Copyright © 2023, The Authors. All rights reserved.

关键词： Alignment

Blind deconvolution for single noisy and blurry image using alternating maximum a posteriori estimation with low rank prior

学校读者我要写书评

暂无评论

Blind deconvolution for single noisy and blurry image using ...

International Symposium on Communications and Information Technologies (ISCIT)

作者： Shijie Sun Huaici Zhao Jinfeng Lv Mingguo Hao Bo Li Key Laboratory of Image Understanding and Computer Vision Shenyang China Shenyang Institute of Automation Chinese Academy of Sciences Shenyang China Key Laboratory of Optical-Electronics Information Processing Chinese Academy of Sciences Shenyang China College of Information Shenyang Institute of Engineering Shenyang China

ISBN: (纸本)9781509041008

The purpose of single image blind deconvolution is to estimate the unknown blur kernel from a single observed blurred image and recover the original sharp image. Such task is severely ill-posed and even more challenging especially in the condition that the noise in the input image cannot be negligible. In this paper, the main problem we focus on is how to effectively apply low rank prior to blind deconvolution. A single noisy and blurry image blind deconvolution algorithm is proposed, using alternating maximum a posteriori (MAP) estimation combined with low rank prior. When estimating the intermediate latent image, low rank prior is used as the constraint that is used for noise suppression of the restored image. The denoised intermediate latent image in turn leads to higher quality blur kernel estimation. These two operations are iterated in this manner to arrive at reliable blur kernel estimation. Extensive experiments show the superiority of the proposed method over state-of-the-art techniques, both qualitatively and quantitatively.

关键词： C# languages

Robust Outlier Rejection for 3D Registration with Variational Bayes

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Jiang, Haobo Dang, Zheng Wei, Zhen Xie, Jin Yang, Jian Salzmann, Mathieu CVLab EPFL Switzerland PCA Lab Key Lab of Intelligent Perception and Systems for High-Dimensional Information of Ministry of Education Jiangsu Key Lab of Image and Video Understanding for Social Security School of Computer Science and Engineering Nanjing University of Science and Technology China

Learning-based outlier (mismatched correspondence) rejection for robust 3D registration generally formulates the outlier removal as an inlier/outlier classification problem. The core for this to be successful is to learn the discriminative inlier/outlier feature representations. In this paper, we develop a novel variational non-local network-based outlier rejection framework for robust alignment. By reformulating the non-local feature learning with variational Bayesian inference, the Bayesian-driven long-range dependencies can be modeled to aggregate discriminative geometric context information for inlier/outlier distinction. Specifically, to achieve such Bayesian-driven contextual dependencies, each query/key/value component in our non-local network predicts a prior feature distribution and a posterior one. Embedded with the inlier/outlier label, the posterior feature distribution is label-dependent and discriminative. Thus, pushing the prior to be close to the discriminative posterior in the training step enables the features sampled from this prior at test time to model high-quality long-range dependencies. Notably, to achieve effective posterior feature guidance, a specific probabilistic graphical model is designed over our non-local model, which lets us derive a variational low bound as our optimization objective for model training. Finally, we propose a voting-based inlier searching strategy to cluster the high-quality hypothetical inliers for transformation estimation. Extensive experiments on 3DMatch, 3DLoMatch, and KITTI datasets verify the effectiveness of our method. Code is available at https://***/Jiang-HB/VBReg. Copyright © 2023, The Authors. All rights reserved.

关键词： Bayesian networks

Spatial group-wise enhance: Improving semantic feature learning in convolutional networks

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Li, Xiang Hu, Xiaolin Yang, Jian Pca Lab Key Lab of Intelligent Percept. and Syst. for High-Dimensional Information of Ministry of Education Jiangsu Key Lab of Image and Video Understanding for Social Security School of Computer Science and Engineering Nanjing University of Science and Technology Momenta Department of Computer Science and Technology Tsinghua University

The Convolutional Neural Networks (CNNs) generate the feature representation of complex objects by collecting hierarchical and different parts of semantic subfeatures. These sub-features can usually be distributed in grouped form in the feature vector of each layer [43, 32], representing various semantic entities. However, the activation of these sub-features is often spatially affected by similar patterns and noisy backgrounds, resulting in erroneous localization and identification. We propose a Spatial Group-wise Enhance (SGE) module that can adjust the importance of each sub-feature by generating an attention factor for each spatial location in each semantic group, so that every individual group can autonomously enhance its learnt expression and suppress possible noise. The attention factors are only guided by the similarities between the global and local feature descriptors inside each group, thus the design of SGE module is extremely lightweight with almost no extra parameters and calculations. Despite being trained with only category supervisions, the SGE component is extremely effective in highlighting multiple active areas with various high-order semantics (such as the dog's eyes, nose, etc.). When integrated with popular CNN backbones, SGE can significantly boost the performance of image recognition tasks. Specifically, based on ResNet50 backbones, SGE achieves 1.2% Top-1 accuracy improvement on the imageNet benchmark and 1.0∼2.0% AP gain on the COCO benchmark across a wide range of detectors (Faster/Mask/Cascade RCNN and RetinaNet). Codes and pretrained models are available at https://***/implus/PytorchInsight. Copyright © 2019, The Authors. All rights reserved.

关键词： Semantics