检索结果-内蒙古大学图书馆

Multi-class feature selection via Sparse Softmax with a discriminative regularization

INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS 2025年第1期16卷 159-172页

作者： Sun, Zhenzhen Chen, Zexiang Liu, Jinghua Yu, Yuanlong HuaQiao Univ Coll Comp Sci & Technol Jimei Ave Xiamen 361021 Fujian Peoples R China HuaQiao Univ Xiamen Key Lab Comp Vis & Pattern Recognit Jimei Ave Xiamen 361021 Fujian Peoples R China Fuzhou Univ Coll Comp & Data Sci Wulong Jiangbei Ave Fuzhou 350108 Fujian Peoples R China

Feature selection plays a critical role in many machine learning applications as it effectively addresses the challenges posed by "the curse of dimensionality" and enhances the generalization capability of trained models. However, existing approaches for multi-class feature selection (MFS) often combine sparse regularization with a simple classification model, such as least squares regression, which can result in suboptimal performance. To address this limitation, this paper introduces a novel MFS method called Sparse Softmax Feature Selection ((SFS)-F-2). (SFS)-F-2 combines a l(2,0)-norm regularization with the Softmax model to perform feature selection. By utilizing the l(2,0)-norm, (SFS)-F-2 produces a more precise sparsity solution for the feature selection matrix. Additionally, the Softmax model improves the interpretability of the model's outputs, thereby enhancing the classification performance. To further enhance discriminative feature selection, a discriminative regularization, derived based on linear discriminate analysis (LDA), is incorporated into the learning model. Furthermore, an efficient optimization algorithm, based on the alternating direction method of multipliers (ADMM), is designed to solve the objective function of (SFS)-F-2. Extensive experiments conducted on various datasets demonstrate that (SFS)-F-2 achieves higher accuracy in classification tasks compared to several contemporary MFS methods.

关键词： Multi-class feature selection L-2,L-0-norm regularization Discriminative regularization Alternating direction method of multipliers

来源：评论

学校读者我要写书评

暂无评论

IFAST: Weakly Supervised Interpretable Face Anti-Spoofing From Single-Shot Binocular NIR Images

引用

IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY 2024年 19卷 9270-9284页

作者： Huang, Jiancheng Zhou, Donghao Liu, Jianzhuang Shi, Linxiao Chen, Shifeng Chinese Acad Sci Shenzhen Inst Adv Technol Shenzhen Key Lab Comp Vis & Pattern Recognit Shenzhen 518055 Peoples R China Chinese Univ Hong Kong Dept Comp Sci & Engn Hong Kong Peoples R China Shenzhen Univ Adv Technol Dept Comp Sci & Control Engn Shenzhen Peoples R China

Single-shot face anti-spoofing (FAS) is a key technique for securing face recognition systems, relying solely on static images as input. However, single-shot FAS remains a challenging and under-explored problem due to two reasons: 1) On the data side, learning FAS from RGB images is largely context-dependent, and single-shot images without additional annotations contain limited semantic information. 2) On the model side, existing single-shot FAS models struggle to provide proper evidence for their decisions, and FAS methods based on depth estimation require expensive per-pixel annotations. To address these issues, we construct and release a large binocular NIR image dataset named BNI-FAS, which contains more than 300,000 real face and plane attack images, and propose an Interpretable FAS Transformer (IFAST) that requires only weak supervision to produce interpretable predictions. Our IFAST generates pixel-wise disparity maps using the proposed disparity estimation Transformer with Dynamic Matching Attention (DMA) blocks. Besides, we design a confidence map generator to work in tandem with a dual-teacher distillation module to obtain the final discriminant results. comprehensive experiments show that our IFAST achieves state-of-the-art performance on BNI-FAS, verifying its effectiveness of single-shot FAS on binocular NIR images. The project page is available at https://***/.

关键词： Face recognition Cameras Estimation Transformers Annotations Feature extraction Costs Face anti-spoofing weakly-supervised learning transformer depth estimation knowledge distillation

来源：评论

学校读者我要写书评

暂无评论

TSNet : Tree structure network for human pose estimation

引用

SIGNAL IMAGE AND VIDEO PROCESSING 2022年第2期16卷 551-558页

作者： Wan, TianJun Luo, YanMin Zhang, Zhiqian Ou, Zhilong Huaqiao Univ Coll Comp Sci & Technol Xiamen 361021 Peoples R China Huaqiao Univ Xiamen Key Lab Comp Vis & Pattern Recognit Xiamen 361021 Peoples R China

Multi-person pose estimation in natural scenes has been a hot topic in the recent years. The prediction speed of the top-down methods is affected by the number of people in the scene, so the bottom-up methods has an advantage in natural scenes. However, the study found that the accuracy of human margin joints (the joints farther from the center of the human, such as wrist and ankle) is always lower than that of the joints that are closer to the center of the human (such as shoulder and hip), and the accuracy gap between joints categories is large. Inspiring from the structural characteristics of human body, this paper proposes a tree structure network (TSNet) for human pose estimation, which divides the joints of the human into several levels according to the characteristics of human body structure, and stepwise predicts the joints from human center to human margin. Combining with the global features, the joint features of the next layer are predicted by extracting the correlation between the joint features of the current layer and the joint features of the previous layer. Therefore, each human joint contains not only the joint information of the current layer and the joint information of the previous layer, but also the background information. The experiment results show that this method can effectively alleviate the uneven precision of joints, and the TSNet can effectively improve the accuracy of lower body joints by setting different activation values for different joints. Extensive experiments on MPII datasets demonstrate the effectiveness of our proposed model and method.

关键词： Tree structure network Different activation values Hierarchy of joint division Multi-person pose estimation

来源：评论

学校读者我要写书评

暂无评论

KV Inversion: KV Embeddings Learning for Text-Conditioned Real Image Action Editing 6th

KV Inversion: KV Embeddings Learning for Text-Conditioned Re...

引用

6th Chinese Conference on pattern recognition and computer vision (PRCV)

作者： Huang, Jiancheng Liu, Yifan Qin, Jin Chen, Shifeng Chinese Acad Sci Shenzhen Inst Adv Technol ShenZhen Key Lab Comp Vis & Pattern Recognit Shenzhen Peoples R China Univ Chinese Acad Sci Beijing Peoples R China

ISBN: (纸本)9789819984282;9789819984299

Text-conditioned image editing is a recently emerged and highly practical task, and its potential is immeasurable. However, most of the concurrent methods are unable to perform action editing, i.e. they can not produce results that conform to the action semantics of the editing prompt and preserve the content of the original image. To solve the problem of action editing, we propose KV Inversion, a method that can achieve satisfactory reconstruction performance and action editing, which can solve two major problems: 1) the edited result can match the corresponding action, and 2) the edited object can retain the texture and identity of the original real image. In addition, our method does not require training the Stable Diffusion model itself, nor does it require scanning a large-scale dataset to perform time-consuming training.

关键词： Real imgae editing Diffusion model Text-to-image generation AIGC

来源：评论

学校读者我要写书评

暂无评论

label-reconstruction-based pseudo-subscore learning for action quality assessment in sporting events

引用

APPLIED INTELLIGENCE 2023年第9期53卷 10053-10067页

作者： Zhang, Hong-Bo Dong, Li-Jia Lei, Qing Yang, Li-Jie Du, Ji-Xiang Huaqiao Univ Dept Comp Sci & Technol Xiamen 361000 Fujian Peoples R China Huaqiao Univ Xiamen Key Lab Comp Vis & Pattern Recognit Xiamen 361000 Fujian Peoples R China Huaqiao Univ Fujian Key Lab Big Data Intelligence & Secur Xiamen 361000 Fujian Peoples R China

Most existing action quality assessment (AQA) methods provide only an overall quality score for the input video and lack an evaluation of each substage of the movement process;thus, these methods cannot provide detailed feedback for users. Moreover, the existing datasets do not provide labels for substage quality assessment. To address these problems, in this work, a new label-reconstruction-based pseudo-subscore learning (PSL) method is proposed for AQA in sporting events. In the proposed method, the overall score of an action is not only regarded as a quality label but also used as a feature of the training set. A label-reconstruction-based learning algorithm is built to generate pseudo-subscore labels for the training set. Moreover, based on the pseudo-subscore labels and overall score labels, a multi-substage AQA model is fine-tuned from the PSL model to predict the action quality score of each substage and the overall score for an athlete. Several ablation experiments are performed to verify the effectiveness of each module. The experimental results show that our approach achieves state-of-the-art performance.

关键词： Action quality assessment Substage quality assessment label reconstruction Pseudo-subscore learning Multi-substage AQA model

来源：评论

学校读者我要写书评

暂无评论

Learning to zoom: Exploiting mixed-scale contextual information for object detection

引用

EXPERT SYSTEMS WITH APPLICATIONS 2025年 264卷

作者： Wang, Boying Ji, Ruyi Zhang, Libo Wu, Yanjun Liu, Jing North Univ China Sch Comp Sci & Technol Shanxi Key Lab Machine Vis & Virtual Real Taiyuan 030051 Peoples R China Chinese Acad Sci Inst Software State Key Lab Comp Sci Beijing 101400 Peoples R China Univ Chinese Acad Sci Beijing 100190 Peoples R China Chinese Acad Sci Inst Automat Natl Lab Pattern Recognit Beijing 101400 Peoples R China

With the development of deep neural networks, object detection has made a significant leap. However, apart from intrinsic intra and inter class differences, the objects in real-world scenarios may encounter diverse scales, blurred appearance or even severe occlusion. Inspired by the human behavior of looking at blurred images, i.e., zooming in/out, we propose a Mixed-Scale Network (dubbed for MSNet) to tackle these issues. Specifically, MSNet employs the zoom strategy to exploit mixed-scale contextual information, which fully unleashes the representation ability of deep neural networks. Firstly, the global feature aggregation module and global feature enhancement module aims at aggregating mixed-scale features from the global perspective, which learns the transform offset of pixels to align the higher-res features contextually. Moreover, the local feature aggregation module enriches the instance-level feature by adaptively aligning the features of different receptive fields. Extensive experiments on MS COCO demonstrated the effectiveness of MSNet, yielding significant improvements of 0.7-2.8 points in AP box over baselines when paired with different detectors, backbones, and schedules. In addition, for pixel-level prediction tasks including semantic segmentation and instance segmentation, the proposed method gains consistent improvements of 0.6-2.7 points in mIoU /AP mask over baselines, which substantiates the effectiveness of MSNet further.

关键词： Deep neural networks Object detection Mixed-scale Contextual information

来源：评论

学校读者我要写书评

暂无评论

Classification of aesthetic natural scene images using statistical and semantic features

引用

MULTIMEDIA TOOLS AND APPLICATIONS 2023年第9期82卷 13507-13532页

作者： Biswas, Kunal Shivakumara, Palaiahnakote Pal, Umapada Lu, Tong Blumenstein, Michael Llados, Josep Indian Stat Inst Comp Vis & Pattern Recognit Unit Kolkata India Univ Malaya Fac Comp Sci & Informat Technol Kuala Lumpur Malaysia Nanjing Univ Natl Key Lab Novel Software Technol Nanjing Peoples R China Univ Technol Sydney Sydney NSW Australia Univ Autonoma Barcelona Comp Vis Ctr Barcelona Spain

Aesthetic image analysis is essential for improving the performance of multimedia image retrieval systems, especially from a repository of social media and multimedia content stored on mobile devices. This paper presents a novel method for classifying aesthetic natural scene images by studying the naturalness of image content using statistical features, and reading text in the images using semantic features. Unlike existing methods that focus only on image quality with human information, the proposed approach focuses on image features as well as text-based semantic features without human intervention to reduce the gap between subjectivity and objectivity in the classification. The aesthetic classes considered in this work are (i) Very Pleasant, (ii) Pleasant, (iii) Normal and (iv) Unpleasant. The naturalness is represented by features of focus, defocus, perceived brightness, perceived contrast, blurriness and noisiness, while semantics are represented by text recognition, description of the images and labels of images, profile pictures, and banner images. Furthermore, a deep learning model is proposed in a novel way to fuse statistical and semantic features for the classification of aesthetic natural scene images. Experiments on our own dataset and the standard datasets demonstrate that the proposed approach achieves 92.74%, 88.67% and 83.22% average classification rates on our own dataset, AVA dataset and CUHKPQ dataset, respectively. Furthermore, a comparative study of the proposed model with the existing methods shows that the proposed method is effective for the classification of aesthetic social media images.

关键词： Scene images Personality traits Person behaviour Color features Deep learning Aesthetic images

来源：评论

学校读者我要写书评

暂无评论

From recognition to Prediction: Leveraging Sequence Reasoning for Action Anticipation

引用

ACM TRANSACTIONS ON MULTIMEDIA compUTING COMMUNICATIONS AND APPLICATIONS 2024年第11期20卷 1-19页

作者： Liu, Xin Hao, Chao Yu, Zitong Yue, Huanjing Yue, Jingyu Tianjin Univ Sch Elect & Informat Engn Tianjin Peoples R China Lappeenranta Lahti Univ Technol LUT Sch Engn Sci Comp Vis & Pattern Recognit Lab Lappeenranta Finland Great Bay Univ Sch Comp & Informat Technol Dongguan Peoples R China

The action anticipation task refers to predicting what action will happen based on observed videos, which requires the model to have a strong ability to summarize the present and then reason about the future. Experience and common sense suggest that there is a significant correlation between different actions, which provides valuable prior knowledge for the action anticipation task. However, previous methods have not effectively modeled this underlying statistical relationship. To address this issue, we propose a novel end- to-end video modeling architecture that utilizes attention mechanisms, named Anticipation via recognition and Reasoning (ARR). ARR decomposes the action anticipation task into action recognition and sequence reasoning tasks and effectively learns the statistical relationship between actions by next action prediction (NAP). In comparison to existing temporal aggregation strategies, ARR is able to extract more effective features from observable videos to make more reasonable predictions. In addition, to address the challenge of relationship modeling that requires extensive training data, we propose an innovative approach for the unsupervised pre-training of the decoder, which leverages the inherent temporal dynamics of video to enhance the reasoning capabilities of the network. Extensive experiments on the Epic-kitchen-100, EGTEA Gaze+, and 50salads datasets demonstrate the efficacy of the proposed methods. The code is available at https://***/linuxsino/ARR.

关键词： Action anticipation Action recognition Sequence reasoning

来源：评论

学校读者我要写书评

暂无评论

Learning Image-Adaptive Lookup Tables With Spatial Awareness for Image Harmonization

引用

IEEE TRANSACTIONS ON CONSUMER ELECTRONICS 2024年第1期70卷 155-166页

作者： Huang, Yi Dong, Yu Zhang, He Huang, Jiancheng Chen, Shifeng Chinese Acad Sci Shenzhen Inst Adv Technol Shenzhen Key Lab Comp Vis & Pattern Recognit Shenzhen 518055 Peoples R China Univ Chinese Acad Sci Shenzhen Inst Adv Technol Beijing 101408 Peoples R China Adobe Inc San Jose CA 95110 USA

Image harmonization aims to address inharmony between the foreground and background of composite images. Current methods mainly adopt the encoder-decoder reconstruction architecture, which restricts their application to high-resolution images due to heavy resource cost. Inspired by Photoshop, whether we can leverage some adjustment tools to apply on composite images to generate harmonized ones. This paper gives an answer "yes", and explores a new viewpoint to address image harmonization problem by lookup tables (LUTs) rather than reconstruct full images, which saves a lot of memory. Specifically, a deep model is utilized to encode the down-sampled version of the input image as style-relevant weights to fuse multiple basic LUTs into image-adaptive ones, which are employed to transform the global color of the foreground in source high-resolution composite image to be harmonious with the background. Besides, a spatial attention module is applied to learn the spatial map for revising preliminary LUTs transformation as well as raising the local spatial awareness of the image-adaptive LUTs. Extensive experiments show that our method achieves competitive performance both qualitatively and quantitatively with a smaller model size.

关键词： Table lookup Transformers Image color analysis Task analysis Costs Software Image reconstruction Image harmonization lookup tables image composition

来源：评论

学校读者我要写书评

暂无评论

Gaussian guided frame sequence encoder network for action quality assessment

引用

compLEX & INTELLIGENT SYSTEMS 2023年第2期9卷 1963-1974页

作者： Li, Ming-Zhe Zhang, Hong-Bo Dong, Li-Jia Lei, Qing Du, Ji-Xiang Huaqiao Univ Sch Comp Sci & Technol Xiamen 361000 Peoples R China Huaqiao Univ Xiamen Key Lab Comp Vis & Pattern Recognit Xiamen 361000 Peoples R China Huaqiao Univ Fujian Key Lab Big Data Intelligence & Secur Xiamen 361000 Peoples R China

Can a computer evaluate an athlete's performance automatically? Many action quality assessment (AQA) methods have been proposed in recent years. Limited by the randomness of video sampling and the simple strategy of model training, the performance of the existing AQA methods can still be further improved. To achieve this goal, a Gaussian guided frame sequence encoder network is proposed in this paper. In the proposed method, the image feature of each video frame is extracted by Resnet model. And then, a frame sequence encoder network is applied to model temporal information and generate action quality feature. Finally, a fully connected network is designed to predict action quality score. To train the proposed method effectively, inspired by the final score calculation rule in Olympic game, Gaussian loss function is employed to compute the error between the predicted score and the label score. The proposed method is implemented on the AQA-7 and MTL-AQA datasets. The experimental results confirm that compared with the state-of-the-art methods, our proposed method achieves the better performance. And detailed ablation experiments are conducted to verify the effectiveness of each component in the module.

关键词： Action quality assessment Frame sequence encoder network Gaussian loss function Regression analysis

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：