Human face recognition based on convolutional neural network and augmented dataset

Lu, Peng; Song, Baoye; Xu, Lin

doi:10.1080/21642583.2020.1836526

Cited by 77 publications

(29 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Finally, classification involves probabilistic comparison of the vectors created from feature extraction to identify photos in the database with face images most similar to the probe photo. Many of the top-performing FR algorithms use deep convolutional neural networks Lu, Song, and Xu, 2021) to perform deep-learning tasks of feature extraction, selection, and classification. 8 Therefore, these models train an algorithm to identify and compare facial features contained in images by comparing thousands or millions of facial images with each other or by using image variations of the same face (i.e., data augmentation; see Wang, Wang, and Lian, 2020;Zhuchkov, 2021).…”

Section: Fr Comparison and Similarity Scoresmentioning

confidence: 99%

Finding a Broadly Practical Approach for Regulating the Use of Facial Recognition by Law Enforcement

2023

View full text Add to dashboard Cite

Technological advances in society often have implications for policing. Specifically, much as fingerprinting and DNA evidence were groundbreaking advances that have been used to solve serious crimes, remote biometric identification, such as facial recognition and object identification, is rapidly developing and being used in policing. This has significant implications for legislators and policymakers, criminal justice officials, and society in general that should be understood, and further developments need to include efforts to maximize the benefits of such technology while limiting the potential risks. This report examines current and near-future uses of facial recognition by law enforcement and identifies the various concerns and related policy domains that could be used to guide public policy. This report will be useful for policymakers, criminal justice officials, and interested members of the public. Justice Policy ProgramRAND Social and Economic Well-Being is a division of the RAND Corporation that seeks to actively improve the health and social and economic well-being of populations and communities throughout the world. This research was conducted in the Justice Policy Program within RAND Social and Economic Well-Being. The program focuses on such topics as access to justice, policing, corrections, drug policy, and court system reform, as well as other policy concerns pertaining to public safety and criminal and civil justice. For more information, email

show abstract

Section: Fr Comparison and Similarity Scoresmentioning

confidence: 99%

Finding a Broadly Practical Approach for Regulating the Use of Facial Recognition by Law Enforcement

2023

View full text Add to dashboard Cite

show abstract

“…In the model training step, a convolutional neural network (CNN) [28][29][30][31] based approach is created for the face recognition task. The Inception-ResNet-v1 [32][33][34], a deep CNN architecture with a combination of Inception block and residual neural network, is deployed as our baseline network.…”

Section: Model Trainingmentioning

confidence: 99%

Face recognition under mask-wearing based on residual inception networks

Moungsouy

Tawanbunjerd

Liamsomboon

et al. 2022

ACI

View full text Add to dashboard Cite

PurposeThis paper proposes a solution for recognizing human faces under mask-wearing. The lower part of human face is occluded and could not be used in the learning process of face recognition. So, the proposed solution is developed to recognize human faces on any available facial components which could be varied depending on wearing or not wearing a mask.Design/methodology/approachThe proposed solution is developed based on the FaceNet framework, aiming to modify the existing facial recognition model to improve the performance of both scenarios of mask-wearing and without mask-wearing. Then, simulated masked-face images are computed on top of the original face images, to be used in the learning process of face recognition. In addition, feature heatmaps are also drawn out to visualize majority of parts of facial images that are significant in recognizing faces under mask-wearing.FindingsThe proposed method is validated using several scenarios of experiments. The result shows an outstanding accuracy of 99.2% on a scenario of mask-wearing faces. The feature heatmaps also show that non-occluded components including eyes and nose become more significant for recognizing human faces, when compared with the lower part of human faces which could be occluded under masks.Originality/valueThe convolutional neural network based solution is tuned up for recognizing human faces under a scenario of mask-wearing. The simulated masks on original face images are augmented for training the face recognition model. The heatmaps are then computed to prove that features generated from the top half of face images are correctly chosen for the face recognition.

show abstract

“…Object detection models are used to automatically detect objects in images and videos (for an overview of the general approach, see [9]). While they are used in many contexts, such as facial pattern recognition [10,11] or object tracking [12], they have not been used to automatically detect objects when evaluating mobile eye-tracking scene videos during an authentic science laboratory course. In this study, we illustrate the use of object detection in mobile eye-tracking.…”

Section: Relevant Literature 21 Object Detectionmentioning

confidence: 99%

Mobile Eye-Tracking Data Analysis Using Object Detection via YOLO v4

Kumari

Ruf

Mukhametov

et al. 2021

Sensors

View full text Add to dashboard Cite

Remote eye tracking has become an important tool for the online analysis of learning processes. Mobile eye trackers can even extend the range of opportunities (in comparison to stationary eye trackers) to real settings, such as classrooms or experimental lab courses. However, the complex and sometimes manual analysis of mobile eye-tracking data often hinders the realization of extensive studies, as this is a very time-consuming process and usually not feasible for real-world situations in which participants move or manipulate objects. In this work, we explore the opportunities to use object recognition models to assign mobile eye-tracking data for real objects during an authentic students’ lab course. In a comparison of three different Convolutional Neural Networks (CNN), a Faster Region-Based-CNN, you only look once (YOLO) v3, and YOLO v4, we found that YOLO v4, together with an optical flow estimation, provides the fastest results with the highest accuracy for object detection in this setting. The automatic assignment of the gaze data to real objects simplifies the time-consuming analysis of mobile eye-tracking data and offers an opportunity for real-time system responses to the user’s gaze. Additionally, we identify and discuss several problems in using object detection for mobile eye-tracking data that need to be considered.

show abstract

Human face recognition based on convolutional neural network and augmented dataset

Cited by 77 publications

References 25 publications

Finding a Broadly Practical Approach for Regulating the Use of Facial Recognition by Law Enforcement

Finding a Broadly Practical Approach for Regulating the Use of Facial Recognition by Law Enforcement

Face recognition under mask-wearing based on residual inception networks

Mobile Eye-Tracking Data Analysis Using Object Detection via YOLO v4

Contact Info

Product

Resources

About