Background: What we think we see consists of models constructed in our brains, which may be constrained, limited and perhaps modified at a cerebral level. Patients may view their mirror image differently to how others and the clini‑ cal team view them. Understanding potential variations in perception between real and mirror images is important in clinical practice. The aims were to assess differences in self-perception between frontal facial and mirrored photo‑ graphs, comparing the results with selfie photographs. Methods: Facial photographs were taken by one investigator under standardized conditions for preclinical and clini‑ cal students. Each student took a selfie photograph at rest and smiling using his/her smartphone. A mirror image was generated for each image. Each student was shown his/her original and mirror image, without being informed which was which. For each pair of images, students were asked to choose which photograph they perceived as more attractive. A set of photographs of a male volunteer was shown to all participants, to choose either the original or mir‑ ror image as the more attractive. Results: Most observers preferred the true image of the volunteer (P < 0.05), which may be evidence that most people prefer the true image of others, which is how they normally view them. Most observers preferred their own original photograph in frontal view at rest and smiling (P< 0.05), but preferred the mirror image of their selfie photo‑ graph at rest and smiling (P < 0.05). Conclusions Significant differences in perceptions of attractiveness between true and mirror-reversed frontal and selfie images were found. Observers preferred their image the way they view themselves in a mirror. The selfie is how other people view an individual. If a selfie is flipped horizontally, that is how an individual sees themselves in a mirror. Most observers preferred the mirror image of their selfie, which is how they would view themselves in a mirror.