Line-field confocal optical coherence tomography (LC-OCT) is an optical technique based on a combination of confocal microscopy and optical coherence tomography, allowing three-dimensional (3D) imaging of the skin in vivo with an isotropic spatial resolution of 1.3 µm and up to 400 µm in depth. The 3D cell-resolved images obtained by LC-OCT provide a considerable amount of information for the description and quantification of the upper skin layers using morphological metrics. This study presents metrics for quantifying the epidermis and uses them to describe the variability of the healthy epidermis between different body sites. These metrics include stratum corneum (SC) and epidermis thicknesses, dermal-epidermal junction (DEJ) undulation, and quantification of the keratinocyte network. An artificial intelligence approach was applied to automate the calculation of the metrics. We were able to fully describe and quantify the epidermis on a panel of eight volunteers on seven body areas including the head, the upper limbs and the trunk. Variations in SC/epidermal thickness and DEJ undulation were observed between body sites: the cheek had the thinnest SC and the least undulated DEJ, while the back of the hand presented the thickest SC. For all sites, the process of keratinocyte maturation was evidenced in vivo. These in vivo 3D quantifications open the door in dermo-cosmetics to understanding skin ageing and monitoring the impact of treatments.