Can a Smile Reveal Your Gender?

Abstract

Automated gender estimation has numerous applications including video surveillance, human computer-interaction, anonymous customized advertisement and image retrieval. Most commonly, the underlying algorithms analyze facial appearance for clues of gender. In this work, we propose a novel approach for gender estimation, based on facial behavior in video-sequences capturing smiling subjects. The proposed behavioral approach quantifies gender dimorphism of facial smiling-behavior and is instrumental in cases of (a) omitted appearance-information (e.g. low resolution due to poor acquisition), (b) gender spoofing (e.g. makeup-based face alteration), as well as can be utilized to (c) improve the performance of appearance-based algorithms, since it provides complementary information. The proposed algorithm extracts spatio-temporal features based on dense trajectories, represented by a set of descriptors encoded by Fisher Vectors. Our results suggest that smile-based features include significant gender-clues. The designed algorithm obtains true gender classification rates of 86.3% for adolescents, significantly outperforming two state-of-the-art appearance-based algorithms (OpenBR and how-old.net), while for adults we obtain true gender classification rates of 91.01%, which is comparably discriminative to the better of these appearance-based algorithms.

Mitigating Bias in Gender, Age and Ethnicity Classification: A Multi-task Convolution Neural Network Approach

“…Cheng et al [15] performed an initial investigation, suggesting that the discriminable features for gender classification for children and adults were significantly different. This was affirmed by Dantcheva et al [16] and Bilinski et al [17].…”

Gender Estimation Based on Smile-Dynamics

“…• Which gender can pose smiles more genuinely? Related work of a holistic smile-based gender estimation algorithm can be found in Biliński et al [9].…”

“…In our related work [9], we have presented a holistic approach for smile-based gender estimation, that extracts spatiotemporal features based on dense trajectories, represented by a set of descriptors encoded by Fisher Vectors. The associated true gender classification rates account for 86.3% for adolescents, and 91.01% for adults.…”

Image-based gender estimation from body and face across distances

“…We note that gender has been predominantly estimated from facial images [10]. Other modalities that have been studied in this context include gait [36], facial smiling behavior [4], [9], or speech [37]. In addition, there are some hybrid approaches fusing algorithms based on different biometric modalities such as face and gait [41], face and fingerprint [24] or face and the shoulder region [23], among others.…”