检索结果-内蒙古大学图书馆

A Survey of Deep Learning for Low-shot object detection

ACM COMPUTING SURVEYS 2024年第5期56卷 1-37页

作者： Huang, Qihan Zhang, Haofei Xue, Mengqi Song, Jie Song, Mingli Zhejiang Univ 38 Zheda Rd Hangzhou Zhejiang Peoples R China

object detection has achieved a huge breakthrough with deep neural networks and massive annotated data. However, current detection methods cannot be directly transferred to the scenario where the annotated data is scarce due to the severe overfitting problem. Although few-shot learning and zero-shot learning have been extensively explored in the field of image classification, it is indispensable to design new methods for object detection in the data-scarce scenario, since object detection has an additional challenging localization task. Low-shot object detection (LSOD) is an emerging research topic of detecting objects from a few or even no annotated samples, consisting of One-shot object Localization (OSOL), few-shot object detection (FSOD), and Zero-shot object detection (ZSOD). This survey provides a comprehensive review of LSOD methods. First, we propose a thorough taxonomy of LSOD methods and analyze them systematically, comprising some extensional topics of LSOD (semi-supervised LSOD, weakly supervised LSOD, and incremental LSOD). Then, we indicate the pros and cons of current LSOD methods with a comparison of their ***, we discuss the challenges and promising directions of LSOD to provide guidance for future works.

关键词： few-shot object detection one-shot object detection zero-shot object detection transfer-learning meta-learning

来源：评论

学校读者我要写书评

暂无评论

Contrastive JS: A Novel Scheme for Enhancing the Accuracy and Robustness of Deep Models

引用

IEEE TRANSACTIONS ON MULTIMEDIA 2023年 25卷 7881-7893页

作者： Xing, Weiwei Yao, Jie Liu, Zixia Liu, Weibin Zhang, Shunli Wang, Liqiang Beijing Jiaotong Univ Sch Software Engn Beijing 100044 Peoples R China Beijing Informat Sci & Technol Univ Sch Informat Management Beijing 100192 Peoples R China Beijing Jiaotong Univ Sch Comp & Informat Technol Beijing 100044 Peoples R China Univ Cent Florida Dept Comp Sci Orlando FL 32816 USA

Deep learning technologies have been applied in various computer vision tasks in recent years. However, deep models suffer performance decay when some unforeseen data are contained in the testing dataset. Although data enhancement techniques can alleviate this dilemma, the diversity of real data is too tremendous to simulate. To tackle this challenge, we study a scheme for improving the robustness and efficiency of the deep network training process in visual tasks. Specifically, first, we build positive and negative sample pairs based on a class-sensitive strategy. Then, we construct a feature-consistent learning strategy based on contrastive learning to constrain the representations of interclass features while paying attention to the intraclass features. To extend the effect of the consistent strategy, we propose a novel contrastive Jensen-Shannon divergence consistency loss (JS loss) to restrict the probability distributions of different sample pairs. The proposed scheme successfully enhances the robustness and accuracy of the utilized model. We validated our approach by conducting extensive experiments in the domains of model robustness and few-shot object detection (FSOD). The results showed that the proposed method achieved remarkable gains over state-of-the-art (SOTA) methods. We obtained a 3.2% average improvement over the best-performing FSOD method.

关键词： Contrastive Learning Deep Model Robustness few-shot object detection Image Augmentation

来源：评论

学校读者我要写书评

暂无评论

FS-OreDet: Feature enhancement and relationship exploration for boosting few-shot object detector of ore images

引用

ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE 2024年第PartE期133卷

作者： Sun, Guodong Cheng, Le Liu, Jinyu Peng, Yuting Xu, Chengming Fu, Yanwei Wu, Bo Zhang, Yang Hubei Univ Technol Sch Mech Engn Wuhan 430068 Peoples R China Beijing Key Lab Proc Automat Min & Met State Key Lab Intelligent Optimized Mfg Mining & M Beijing 102628 Peoples R China Nanjing Univ Natl Key Lab Novel Software Technol Nanjing 210023 Peoples R China Fudan Univ Sch Data Sci Shanghai 200433 Peoples R China Chinese Acad Sci Shanghai Adv Res Inst Shanghai 201210 Peoples R China Hubei Univ Technol Hubei Key Lab Modern Mfg Qual Engn Wuhan 430068 Peoples R China

In the ore beneficiation process, large block detection is necessary to ensure production safety. This typically involves identifying oversized ore on the conveyor belt and preventing material blockage accidents in the transfer buffer bin between the ore feeding belt and the ore receiving belt. Methods based on deep learning can learn to construct complex features from a large amount of data, but they also require a large number of handmade datasets for training. Although the existing few shot detection methods for ore images reduce the cost of manual labeling, the corresponding detection performance is insufficient. This article mainly explores how to improve the performance of the detector under the ore image detection task in the case of few labeled images. First, a shot enhancement block is proposed to enhance the valuable foreground information for higher -quality support features. Subsequently, we present a dual -attention region proposal network that effectively leverages support features to enhance the precision of generating candidate proposals. Finally, we propose a lightweight multi -relational detector to effectively evaluate the relationship between query and support proposals, leading to a substantial enhancement in guidance performance. The proposed few -shot object detector (FS-OreDet) achieves the best detection results with state-of-the-art methods with an average precision ( AP ) of 55.1, a speed of 57 frames per second ( FPS ), and a model size of only 17 MB . Furthermore, our framework adeptly captures the feature information of ore images with substantial data. The detector's accuracy achieves a significant improvement of 14% in AP . Compared with general object detectors, the performance of the detector ranks first and meets the requirements for outdoor scene deployment.

关键词： Ore beneficiation process Material blockage few-shot object detection Feature enhancement Relationship exploration

来源：评论

学校读者我要写书评

暂无评论

Meta-DETR: Image-Level few-shot detection With Inter-Class Correlation Exploitation

引用

IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 2023年第11期45卷 12832-12843页

作者： Zhang, Gongjie Luo, Zhipeng Cui, Kaiwen Lu, Shijian Xing, Eric P. Nanyang Technol Univ Sch Comp Sci & Engn Singapore 639798 Singapore Carnegie Mellon Univ Sch Comp Sci Pittsburgh PA 15213 USA Mohamed bin Zayed Univ Artificial Intelligence Abu Dhabi U Arab Emirates

few-shot object detection has been extensively investigated by incorporating meta-learning into region-based detection frameworks. Despite its success, the said paradigm is still constrained by several factors, such as (i) low-quality region proposals for novel classes and (ii) negligence of the inter-class correlation among different classes. Such limitations hinder the generalization of base-class knowledge for the detection of novel-class objects. In this work, we design Meta-DETR, which (i) is the first image-level few-shot detector, and (ii) introduces a novel inter-class correlational meta-learning strategy to capture and leverage the correlation among different classes for robust and accurate few-shot object detection. Meta-DETR works entirely at image level without any region proposals, which circumvents the constraint of inaccurate proposals in prevalent few-shot detection frameworks. In addition, the introduced correlational meta-learning enables Meta-DETR to simultaneously attend to multiple support classes within a single feedforward, which allows to capture the inter-class correlation among different classes, thus significantly reducing the misclassification over similar classes and enhancing knowledge generalization to novel classes. Experiments over multiple few-shot object detection benchmarks show that the proposed Meta-DETR outperforms state-of-the-art methods by large margins. The implementation codes are publicly available at https://***/ZhangGongjie/Meta-DETR.

关键词： Proposals Correlation object detection Feature extraction Detectors Task analysis Transformers few-shot learning meta-learning few-shot object detection class correlation

来源：评论

学校读者我要写书评

暂无评论

Yolo-sd: simulated feature fusion for few-shot industrial defect detection based on YOLOv8 and stable diffusion

引用

INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS 2024年第10期15卷 4589-4601页

作者： Wen, Yihao Wang, Li Taiyuan Univ Technol Coll Comp Sci & Technol Coll Data Sci 79 Yingze West St Taiyuan 030024 Shanxi Province Peoples R China

Defect detection from images, an important application in the development of the industrial internet, has been gaining increasing attention due to its close relationship with product quality in industrial production. However, two major challenges in defect detection persist: (1) Limited availability of datasets. Deep learning-based models typically require large-scale training sets to achieve satisfactory detection results. (2) Insufficient image quality and low detection accuracy. When many object detection methods are applied to industrial defect detection, they often exhibit poor performance in handling unclear boundaries, complex backgrounds, noise, and textures. In this study, we propose an advanced defect detection method based on YOLO and Stable Diffusion (YOLO-SD). For the few-shot dataset, a controllable generation module is designed, that integrates CLIP, LoRA, and ControlNet based on Stable Diffusion. Among them, CLIP text inversion can generate the most suitable prompt words from the defect dataset, providing prompt input for Stable Diffusion. LoRA can intervene and adjust the image style of Stable Diffusion by training on the defect dataset in a fine-tuning way. ControlNet obtains boundary and depth maps through HED and Midas. For the insufficient image quality and low detection accuracy, an improved YOLO model with an attention-based Fusion Simulated Feature module (FSF) is built that extracts defect features of the original images and generated images, which provides richer semantic information to improve the detection accuracy. At the same time, in order to make the model lightweight, we introduce a test optimization strategy to improve the model training process. Extensive experiments on the NEU-DET steel defect dataset show that the images generated by our method can expand the dataset to train the model and achieve a certain improvement in defect detection.

关键词： YOLO NEU-DET Stable Diffusion few-shot object detection

来源：评论

学校读者我要写书评

暂无评论

Online visual end-to-end detection monitoring on surface defect of aluminum strip under the industrial few-shot condition

引用

JOURNAL OF MANUFACTURING SYSTEMS 2023年 70卷 31-47页

作者： Ma, Zhuxi Li, Yibo Huang, Minghui Deng, Nanzhou Cent South Univ Sch Mech & Elect Engn Changsha 410083 Hunan Peoples R China Cent South Univ Light Alloy Res Inst Changsha 410083 Hunan Peoples R China

Surface defect detection systems based on deep learning are employed in the manufacturing system, and their good detection performance largely relies on abundant annotated data. Nevertheless, industrial datasets are often difficult to obtain, which hinders the development of defect detection systems to some extent. In an effort to address the issue that the scarcity of training data leads to poor performance, this paper proposes a novel end to end few-shot detection method for industrial real-time detection on aluminum strips. The meta-learning theory is introduced into the multi-scale structure of the YOLOv4 framework, which means that multi-scale meta-feature maps are extracted through the backbone network and each meta-feature map is recalibrated by reweighted vectors to strengthen class-specific features. Subsequently, a multi-scale prediction module with a new parsing strategy is designed to locate and classify the defects at different scales for the reweighted meta-features. All experiments are performed using data collected from the aluminum strip production line in a cold rolling workshop. In addition, the dataset required for training is constructed based on the few-shot training strategy. Pre-training is conducted on a large amount of labeled data from base classes in advance, and the pre-trained weights are then loaded to train the model on a small amount of labeled data from all classes to rapidly deploy the model in multi-task scenarios. The proposed method shows excellent performance, improving the accuracy for novel classes by more than 14.6%, which enhances the industrial practicality of detection systems.

关键词： Surface defects detection Aluminum alloy strip few-shot object detection Meta learning Multi-scale Intelligence manufacturing

来源：评论

学校读者我要写书评

暂无评论

ITFD: an instance-level triplet few-shot detection network under weighted pair-resampling

引用

APPLIED INTELLIGENCE 2023年第19期53卷 22728-22742页

作者： Chen, Xin Peng, Chaoyong Qiu, Chunrong Luo, Lin Huang, Deqing Liu, Ziyi Southwest Jiaotong Univ Sch Phys Sci & Technol 111Sect 1North Erhuan Rd Chengdu 610036 Sichuan Peoples R China Southwest Jiaotong Univ Inst Syst Sci & Technol Sch Elect Engn 999 Xian Rd Chengdu 611756 Sichuan Peoples R China

few-shot object detection has been widely applied in industrial applications, endangered detection, tumor lesion detection, etc. Although many excellent few-shot detection models have been proposed recently, intra-class & inter-class confusion and low activation of novel classes still keep few-shot detection challenging. In this paper, we propose a novel few-shot detection model ITFD, in which a weighted pair-resampling method improves the localization efficiency of the novel-class and a hard triplet loss reduces intra-class & inter-class confusion are contained. Extensive experiments have shown that our model achieves 3.6%, 2.6%, and 3.7% average nAP50 improvement on novel-class setup 1,2,3 of PASCAL VOC compared to the same one-time fine-tuning type of models. Besides, to verify the effectiveness of our model in practical application, we established two train component detection datasets. Our model achieves state-of-the-art performance on both datasets with an average nAP50 improvement of 7% and 4.8% respectively.

关键词： few-shot object detection few-shot learning Contrastive learning Sampling methods

来源：评论

学校读者我要写书评

暂无评论

IRA-FSOD: Instant-Response and Accurate few-shot object Detector

引用

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY 2023年第11期33卷 6912-6923页

作者： Huang, Junying Cao, Junhao Lin, Liang Zhang, Dongyu Sun Yat Sen Univ Dept Comp Sci & Engn Guangzhou 510275 Peoples R China

Aiming at recognizing and localizing objects of novel categories with just a few reference samples, few-shot object detection (FSOD) is quite a challenging task. Previous works rely heavily on the fine-tuning process to transfer their models to the novel categories. They are flawed in the real application since the fine-tuning process is time-consuming and it suffers from serious deterioration on the low-quality support set. Based on the observation, this paper proposes an instant-response and accurate few-shot object detector (IRA-FSOD) that can detect the objects from novel categories without fine-tuning. We carefully analyze the limitations of widely-used Faster R-CNN and transform it to IRA-FSOD. Specifically, we first propose a novel semi-supervised Region Proposal Network (SS-RPN) module and a switch classifier module to precisely recognize the potential foreground instances from novel categories without fine-tuning. Moreover, we introduce two explicit inference strategies into the localization module, including explicit localization score and semi-explicit box regression, to alleviate over-fitting towards the base categories. Extensive experiments demonstrates that the proposed IRA-FSOD not only accomplish few-shot object detection with the instant-response, but also reaches state-of-the-art performance under various FSOD protocols and settings.

关键词： Deep learning object detection few-shot object detection few-shot learning instant-response

来源：评论

学校读者我要写书评

暂无评论

Multi-task Self-supervised few-shot detection 6th

Multi-task Self-supervised Few-Shot Detection

引用

6th Chinese Conference on Pattern Recognition and Computer Vision (PRCV)

作者： Zhang, Guangyong Duan, Lijuan Wang, Wenjian Gong, Zhi Ma, Bian Beijing Univ Technol Fac Informat Technol Beijing 100124 Peoples R China Beijing Key Lab Trusted Comp Beijing 100124 Peoples R China Natl Engn Lab Crit Technol Informat Secur Classif Beijing 100124 Peoples R China

ISBN: (纸本)9789819985548;9789819985555

few-shot object detection involves detecting novel objects with only a few training samples. But very few samples are difficult to cover the bias of the new class in the deep model. To address the issue, we use self-supervision to expand the coverage of samples to provide more observation angles for new classes. In this paper, we propose a multi-task approach that combines self-supervision with few-shot learning to exploit the complementarity of these two domains. Specifically, our self-supervision as an auxiliary task to improve the detection performance of the main task of few-shot learning. Moreover, in order to make self-supervision more suitable for few-shot object detection, we introduce the denoising module to expand the positive and negative samples and the team module for precise positioning. The denoising module expands the positive and negative samples and accelerate model convergence using contrastive denoising training methods. The team module utilizes location constraints for precise localization to improve the accuracy of object detection. Our experimental results demonstrate the effectiveness of our method on the few-shot object detection task on the PASCAL VOC and COCO datasets, achieving promising results. Our results highlight the potential of combining self-supervision with few-shot learning to improve the performance of object detection models in scenarios where annotated data is limited.

关键词： few-shot object detection Self-supervised learning End-to-End Detector

来源：评论

学校读者我要写书评

暂无评论

Robust Temporally-Coherent Strategy for few-shot Video Instance Segmentation 29

Robust Temporally-Coherent Strategy for Few-shot Video Insta...

引用

IEEE International Conference on Image Processing (ICIP)

作者： Wang, Qiuyue Zhang, Songyang He, Xuming ShanghaiTech Univ Shanghai Peoples R China

ISBN: (数字)9781665496209

ISBN: (纸本)9781665496209

Traditional video instance segmentation (VIS) aims to detect, segment, and track object instances from a known class set in videos. In real-world applications, however, video instance segmentation typically need to cope with novel-class instances and to fast adapt with a few labeled videos. In this work, we aim to tackle the task of few-shot video instance segmentation (FVIS), which is challenging due to large variations in object appearance and motion. We propose a robust temporally coherent strategy, termed as VTFA, based on a two-stage fine-tuning approach. VTFA enforces the instance segmentation of novel classes to be temporally smooth and reduces the classification bias between novel and base classes. The proposed Memory-aware Temporal Context Encoding Module (MTCE) in VTFA encodes the temporal context information, which contributes to the consistency in the final predictions. We also propose a loss named Instance-level Pair-wise Contrastive (IPC) Loss on both the novel and base classes to enhance the robustness of instance classification. To validate our method, we develop a YouTube-VIS-FS benchmark to compare our method with several baselines. The experimental evaluation shows that our strategy is superior or competitive to those strong baselines.

关键词： few-shot Video Instance Segmentation few-shot object detection few-shot Learning

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：