Effective Unconstrained Face Recognition by Combining Multiple Descriptors and Learned Background Statistics

By Lior Wolf, Tal Hassner, Yaniv Taigman. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011

368 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/133f01aec1534604d184d56de866a4bd531dac87

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

PCANet: A Simple Deep Learning Baseline for Image Classification? (2014, methodology): “using LFW-a [32], an aligned version of LFW images using the commercial alignment system of face.”
Information Bottleneck Domain Adaptation with Privileged Information for Visual Recognition (2016, methodology): “Gender recognition: We evaluate the proposed approach also for gender recognition where we use the RGB-D face dataset EURECOM [27] as source domain, and the RGB dataset Labeled Faces in the Wild-a (LFW-a) [48] as target domain.”
Recognizing RGB Images by Learning from RGB-D Data (2014, methodology): “The images in the LFW-a dataset have been aligned according to the eye-positions (see [30] for details).”
Global-Local Face Upsampling Network (2016, methodology): “The uncontrolled setting was conducted using an aligned version of the Labeled Faces in the Wild dataset [42], called Labeled Faces in the Wild-a (LFWa) [43].”
Effective face frontalization in unconstrained images (2014, background): “…that pool information over large image regions, thereby accounting for possible misalignments due to pose changes (e.g., [20, 25, 33, 36]), by improving 2D face alignment accuracy [14, 15, 35], or by using massive face collections to learn pose-robust representations [13, 26, 27, 41].”
Face Recognition via Archetype Hull Ranking (2013, methodology): “The roughly aligned images of LFW are from the LFW-a dataset [29].”
Probabilistic Elastic Matching for Pose Variant Face Verification (2013, methodology): “In recent years, we have witnessed more and more research efforts on face recognition under uncontrolled settings [21, 14, 18, 29, 7].”
Trunk-Branch Ensemble Convolutional Neural Networks for Video-Based Face Recognition (2016, background): “the variable volume of video frames and complex facial variations that exist in real-world scenarios [26]. Extracting high-quality face representations has always been a core task in face recognition [27], [28]. In contrast to still face images, video frames usually suffer from severe image blur bec”
Age and Gender Estimation of Unfiltered Faces (2014, background): “These were selected due to their successful application to face recognition problems [60],”
Augmenting CRFs with Boltzmann Machine Shape Priors for Image Labeling (2013, methodology): “Although some better automatic alignments of these images exist, such as the LFW-a data set [31], LFW-a does not contain color information, which is important for our application.”
Motion Interchange Patterns for Action Recognition in Unconstrained Videos (2012, positive): “Combining this line of reasoning with the effectiveness of LBP [20] in encoding texture for recognition tasks, LTP [22], and also patch-LBP [24, 25], use SSDs as a basic building block for defining local patterns.”
Multi-Directional Multi-Level Dual-Cross Patterns for Robust Face Recognition (2014, background): “Index Terms—Face recognition, face image descriptors, face image representation F”
Face Recognition Across Non-Uniform Motion Blur, Illumination, and Pose (2015, methodology): “We use the LFW-aligned version [41] of the database to report our scores with Chi square distance as our similarity measure.”
Emotion Recognition in the Wild via Convolutional Neural Networks and Mapped Binary Patterns (2015, methodology): “Used with Support Vector Machines (SVM) classifiers, LBP code histograms have been key to the success of face recognition systems [41, 42, 14].”
Median Robust Extended Local Binary Pattern for Texture Classification (2016, methodology): “Wolf et al. [32] proposed Three Patch LBP (TPLBP) and Four Patch LBP (FPLBP) using averaged patch difference magnitudes.”
Learning Pose-Aware Models for Pose-Invariant Face Recognition in the Wild (2019, background): “Frame and image pooling: A number of previous efforts proposed different ways of pooling feature descriptors [51], [52].”