AFFACT: Alignment-free facial attribute classification technique

Abstract

Facial attributes are soft-biometrics that allow limiting the search space, e.g., by rejecting identities with nonmatching facial characteristics such as nose sizes or eyebrow shapes. In this paper, we investigate how the latest versions of deep convolutional neural networks, ResNets, perform on the facial attribute classification task. We test two loss functions: the sigmoid cross-entropy loss and the Euclidean loss, and find that for classification performance there is little difference between these two. Using an ensemble of three ResNets, we obtain the new state-of-the-art facial attribute classification error of 8.00 % on the aligned images of the CelebA dataset. More significantly, we introduce the Alignment-Free Facial Attribute Classification Technique (AFFACT), a data augmentation technique that allows a network to classify facial attributes without requiring alignment beyond detected face bounding boxes. To our best knowledge, we are the first to report similar accuracy when using only the detected bounding boxes -rather than requiring alignment based on automatically detected facial landmarks -and who can improve classification accuracy with rotating and scaling test images. We show that this approach outperforms the CelebA baseline on unaligned images with a relative improvement of 36.8 %.

Auglabel: Exploiting Word Representations to Augment Labels for Face Attribute Classification

“…In addition, our method is competitive to existing state-of-the-art methods. As stated before, our method is not directly comparable to the methods: AFFACT [13] and of Sun et al [29] as these methods rely on multiple number of CNNs. Moreover, our method is generic and complementary in nature and can be applied in any architecture of CNNs.…”

“…Aug. + Dropout) 90.9% Proposed (Geo. Aug. + Dropout + AugLabel) 91.2% Face Tracer [17] 81.1% LeNet+ANet [20] 87.3% MOON [26] 90.9% Walk and Learn [30] 88.7% MCNN [14] 91.2% AFFACT [13] 91.5% Kalayeh et al [15] 91.2% Sun et al [29] 91.6% Fig. 3 shows qualitative visualisations of some of the predictions made by the baseline (Geo.…”

A Survey of Deep Facial Attribute Analysis

“…More importantly, this approach can handle in-the-wild input images with complex illumination and occlusions, and no extra cropping and aligning operations are needed. Ding et al [18] propose a cascade network to locate face regions according to different attributes and perform FAE simultaneously with no need to align faces [31]. Li et al [63] design an AFFAIR network for learning a hierarchy of spatial transformations and predicting facial attributes without landmarks.…”

FlowSAN: Privacy-Enhancing Semi-Adversarial Networks to Confound Arbitrary Face-Based Gender Classifiers

“…The experiments designed in this section assess how well the proposed models are able to confound gender classifiers that were unseen during training. These six gender classifiers include three models that were already trained: a commercial-of-the-shelf gender classifier (G-COTS), IntraFace [55], AFFACT [56], and three CNN models built in-house, which we refer to as CNN-1, CNN-2 (trained using MORPH-train and LFW, respectively), and CNN-3 (trained on the union of MORPH-train and LFW). Note that these three CNN models have shown a similar level of performance on the original test-sets, compared to the other three pre-trained gender predictors.…”