Pixel-Level Contrastive Pretrainer for Industrial Image Representation
文献类型：期刊论文
| 作者 | Zhu, Bingke1 ; Chen, Yingying1; Tang, Ming1; Wang, Jinqiao1,2,3,4
|
| 刊名 | IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
 |
| 出版日期 | 2024 |
| 卷号 | 73页码:13 |
| 关键词 | Task analysis Feature extraction Anomaly detection Transformers Production Inspection Convolutional neural networks Ind-2M industrial image representation Pixel-level COntrastive (PiCO) pixel-level contrastive learning pretrainer |
| ISSN号 | 0018-9456 |
| DOI | 10.1109/TIM.2024.3353860 |
| 通讯作者 | Chen, Yingying(yingying.chen@nlpr.ia.ac.cn) |
| 英文摘要 | Industrial quality inspection aims to identify defective parts in industrial production processes. Commonly used methods for industrial quality inspection rely on feature representations that have been pretrained on natural image datasets, such as ImageNet. However, these pretrained models are not specifically tailored for industrial scenarios and therefore do not transfer well to downstream industrial tasks. In this study, we have curated a large-scale industrial production dataset called Ind-2M, which is specifically collected from industrial scenarios. This dataset serves to enhance the industrial representation of pretraining models. Additionally, we propose a Pixel-level COntrastive (PiCO) pretrainer for industrial image representation. PiCO not only improves the global industrial representation through industrial production classification, but also enhances the local industrial representation through pixel-level self-supervision. Experimental results demonstrate that PiCO effectively transfers to downstream industrial tasks, such as multilabel defect classification and anomaly detection, outperforming existing pretrained methods. We hope PiCO can initiate a new paradigm for industrial image pretraining. |
| 资助项目 | National Key Research and Development Program of China |
| WOS研究方向 | Engineering ; Instruments & Instrumentation |
| 语种 | 英语 |
| WOS记录号 | WOS:001167387300004 |
| 出版者 | IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC |
| 资助机构 | National Key Research and Development Program of China |
| 源URL | [http://ir.ia.ac.cn/handle/173211/57848]  |
| 专题 | 紫东太初大模型研究中心_大模型计算 |
| 通讯作者 | Chen, Yingying |
| 作者单位 | 1.Chinese Acad Sci, Inst Automat, Fdn Model Res Ctr, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China 3.AI Res, Wuhan 430073, Peoples R China 4.Peng Cheng Lab, Shenzhen, Peoples R China |
| 推荐引用方式 GB/T 7714 | Zhu, Bingke,Chen, Yingying,Tang, Ming,et al. Pixel-Level Contrastive Pretrainer for Industrial Image Representation[J]. IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT,2024,73:13. |
| APA | Zhu, Bingke,Chen, Yingying,Tang, Ming,&Wang, Jinqiao.(2024).Pixel-Level Contrastive Pretrainer for Industrial Image Representation.IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT,73,13. |
| MLA | Zhu, Bingke,et al."Pixel-Level Contrastive Pretrainer for Industrial Image Representation".IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT 73(2024):13. |
入库方式： OAI收割
来源：自动化研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。