RMFD (Real-World Masked Face Dataset)

Real-World Masked Face Dataset (RMFD) 是用于蒙面人脸检测的大型数据集。为了有效防止 COVID-19 病毒的传播，在冠状病毒流行期间几乎每个人都戴着口罩。这几乎使得传统的人脸识别技术在很多情况下都失效了，比如小区门禁、人脸门禁、人脸考勤、火车站人脸安检等。因此，提高现有人脸识别的识别性能非常迫切。蒙面脸上的技术。当前大多数先进的人脸识别方法都是基于深度学习设计的，它依赖于大量的人脸样本。然而，目前还没有公开可用的蒙面人脸识别数据集。为此，本工作提出了三种蒙面人脸数据集，包括蒙面人脸检测数据集（MFDD）、真实蒙面人脸识别数据集（RMFRD）和模拟蒙面人脸识别数据集（SMFRD）。其中，据我们所知，RMFRD 是目前世界上最大的真实世界蒙面人脸数据集。这些数据集可供工业界和学术界免费使用，基于这些数据集可以开发各种蒙面应用程序。我们开发的多粒度蒙面人脸识别模型准确率达到95%，超过了业界报道的结果。我们的数据集位于：https://github.com/X-zhangyang/Real-World-Masked-Face-Dataset。

Real-World Masked Face Dataset (RMFD) is a large-scale dataset dedicated to masked face detection. To effectively curb the spread of COVID-19, nearly all people wore masks during the coronavirus pandemic, which rendered traditional face recognition technologies ineffective in numerous scenarios such as community access control, facial recognition door locks, facial attendance systems, and facial security checks at railway stations. Therefore, it is extremely urgent to enhance the recognition performance of existing face recognition technologies on masked faces. Most current state-of-the-art face recognition methods are designed based on deep learning, which rely on massive face samples. However, no publicly available masked face recognition datasets have been released so far. To address this gap, this work proposes three masked face datasets, including Masked Face Detection Dataset (MFDD), Real-World Masked Face Recognition Dataset (RMFRD), and Simulated Masked Face Recognition Dataset (SMFRD). To the best of our knowledge, RMFRD is the largest real-world masked face dataset globally currently. These datasets are freely accessible for both industrial and academic communities, and various masked face applications can be developed based on them. The multi-granularity masked face recognition model we developed achieves an accuracy of 95%, which outperforms the results reported in the industry. Our dataset is available at: https://github.com/X-zhangyang/Real-World-Masked-Face-Dataset.

搜集汇总

背景与挑战

背景概述

RMFD是一个用于蒙面人脸检测的大型真实世界数据集，由武汉大学于2020年发布，旨在应对COVID-19疫情期间佩戴口罩导致传统人脸识别技术失效的挑战。该数据集包含三个子集，其中真实蒙面人脸识别数据集（RMFRD）是目前世界上最大的真实世界蒙面人脸数据集，可用于开发蒙面人脸识别应用，基于该数据集的多粒度模型准确率达到95%。