Face Space Representations in Deep Convolutional Neural Networks

doi:10.1016/j.tics.2018.06.006

Paper

Face Space Representations in Deep Convolutional Neural Networks

Published Sep 1, 2018 · A. O’toole, C. Castillo, Connor J. Parde

Trends in Cognitive Sciences

Q1 SJR score

106

Citations

6

Influential Citations

Full text

Semantic Scholar

Abstract

Abstract removed due to Elsevier request; this does not indicate any issues with the research. Click the full text link above to read the abstract and view the original source.

Highly Cited

Study Snapshot

Key takeawayDeep convolutional neural networks (DCNNs) provide a new class of visual representation for faces, allowing for generalized face recognition across variations in viewpoint, illumination, expression, and appearance.

PopulationOlder adults (50-71 years)

Sample size24

MethodsObservational

OutcomesBody Mass Index projections

ResultsSocial networks mitigate obesity in older groups.

Sign up to use Study Snapshot

Consensus is limited without an account. Create an account or sign in to get more searches and use the Study Snapshot.

Create an account

Paper

Face Space Representations in Deep Convolutional Neural Networks

Sign up to use Study Snapshot

References

Face recognition accuracy of forensic examiners, superrecognizers, and face recognition algorithms

Collaboration among humans and machines, particularly forensic facial examiners and superrecognizers, can significantly improve face identification accuracy in forensic applications.

Are We Face Experts?

We are, in effect, familiar face experts, as we show limited expertise with unfamiliar faces for identity recognition.

Seeing it all: Convolutional network layers map the function of the human visual system

Convolutional networks can accurately model human brain activity during image viewing, capturing universal representations of brain function across various experimental paradigms.

An All-In-One Convolutional Neural Network for Face Analysis

Our single deep convolutional neural network (CNN) effectively performs simultaneous face detection, alignment, pose estimation, genderrecognition, smile detection, age estimation, and face recognition tasks, achieving state-of-the-art results for most of these tasks.

Explicit information for category-orthogonal object properties increases along the ventral stream

The inferior temporal visual area encodes all category-orthogonal object properties more explicitly and better predicts human performance patterns than earlier ventral stream areas.

Citations