State-of-the-art deep learning-based Head Pose Estimation (HPE) techniques have reached spectacular performance on High-Resolution (HR) face images. However, they still fail to achieve expected performance on low-resolution images at large scales. This work presents an end-to-end HPE framework assisted by a Face Super-Resolution (FSR) algorithm. The proposed FSR model is specifically guided to enhance the HPE performance rather than considering FSR as an independent task. To this end, we utilized a Multi-Stage Generative Adversarial Network (MSGAN) which benefit from a pose-aware adversarial loss and head pose estimation feedback to generate super-resolved images that are properly aligned for HPE. Also, we propose a degradation strategy rather than simple down-sampling approach to mimic the diverse properties of real-world Low-Resolution (LR) images. We evaluate the performance of our proposed method on both synthetic and real-world LR datasets and show the superiority of our approach in both visual and HPE metrics on the AFLW2000, BIWI, and WiderFace Datasets.INDEX TERMS Head pose estimation (HPE), face super-resolution (FSR), multi-stage generative adversarial networks (MSGAN), low-resolution (LR) face images.

Currently available face datasets mainly consist of a large number of high-quality and a small number of lowquality samples. As a result, a Face Recognition (FR) network fails to learn the distribution of low-quality samples since they are less frequent during training (underrepresented). Moreover, current state-of-the-art FR training paradigms are based on the sample-to-center comparison (i.e., Softmax-based classifier), which results in a lack of uniformity between train and test metrics. This work integrates a quality-aware learning process at the sample level into the classification training paradigm (QAFace). In this regard, Softmax centers are adaptively guided to pay more attention to low-quality samples by using a quality-aware function. Accordingly, QAFace adds a quality-based adjustment to the updating procedure of the Softmax-based classifier to improve the performance on the underrepresented low-quality samples. Our method adaptively finds and assigns more attention to the recognizable low-quality samples in the training datasets. In addition, QAFace ignores the unrecognizable low-quality samples using the feature magnitude as a proxy for quality. As a result, QAFace prevents class centers from getting distracted from the optimal direction. The proposed method is superior to the state-ofthe-art algorithms in extensive experimental results on the CFP-FP, LFW, CPLFW, CALFW, AgeDB, IJB-B, and IJB-C datasets.

No abstract

No abstract