The smell of us – crowdsourcing human body odor evaluation - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Human Computation - A Transdisciplinary Journal Année : 2016

The smell of us – crowdsourcing human body odor evaluation

Résumé

Human body odor is produced when sweat-secreted compounds are metabolized by bacteria present on the skin. The resulting volatile mixture is often negatively perceived, motivating the use of personal cosmetic deodorants. Yet body odor may also be positively perceived in some contexts, and is proposed to play a role in sexual attraction, kin identification and social bonding. Because only human smellers can report the hedonic qualities of body odor, their persceptions are a valualbe complement to modern GC-MS-based quantitative chemical measurements. Here we present a crowdsourcing framework that engages volunteer smellers to characterize human sweat samples. Our approach seeks to reward both the sweat donor and the smeller with a web-based graphical interface that is informative, interesting, and fun. 300 samples from 87 individual donors were scored by 93 smellers for intensity, pleasantness, and a variety of odor descriptors. Body odor intensity and pleasantness were determined to vary with age, gender, and self-reported deodorant use. Counterintuitively, deodorant use showed no effect on the perceived intensity of body odor, and was associated with a decrease in the perceived pleasantness. From these data, we determine the precision and dynamic range of the volunteer nose as a body odor evaluation instrument and estimate the scale of crowdsourcing effort that would be required to precisely quantify the public perception of body odors.

Dates et versions

hal-01672595 , version 1 (26-12-2017)

Identifiants

Citer

Marguerite Benony, Marianne Cardon, Arnaud Ferré, Jean Coquet, Nathan Foulquier, et al.. The smell of us – crowdsourcing human body odor evaluation. Human Computation - A Transdisciplinary Journal, 2016, 3 (1), pp.161-197. ⟨10.15346/hc.v3i1.9⟩. ⟨hal-01672595⟩
423 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More