Elizabeth is an apparel company that is built around the concept of creating a collection of unique children’s apparel that encourages self-expression. The company was founded in January 2016. Their principal office is located in Taipei, Taiwan. The brand was initially created on the foundation of the Fairyware Couture collection. Each dress in the collection is delicately handmade taking over 4 hours of labor. Demand continued to increase for the collection leading the brand to venture into Ready to Wear and Accessories. Elizabeth had successfully found a demand in the children’s wear market. Elizabeth currently has a total of three different collections with more on the way. The first collection launched in May 2016 through the retail channel of a pop up store. The launch was an extreme success with most of the inventory being depleted within the first week. In addition to several large press publications promoting the brand and the new collection for several weeks leading up to the opening and for several weeks after.

The future of the fashion industry is expected to be online, thus a significant amount of research is being conducted in the field of fashion image analysis. Currently, a task that places a heavy workload on online stores is manually tagging new garments, including attributes such as category, color, pattern, or style. To this end, extensive research has targeted the automatic prediction of clothing categories and attributes, achieving promising results. Nevertheless, no previous study has been found in the literature that specifically reflects the performance of clothing attribute recognition with children’s clothing. This work intends to fill this gap and effectively present, in the same fashion analysis task, how a model trained in adult fashion performs over a model trained exclusively in children’s fashion. When examining the global understanding of children’s fashion apparel, the experiments exhibit that the best performance is obtained when leveraging the domain knowledge of adult fashion, specifically from the iMaterialist dataset, wherein the best model a difference in the overall performance of about 3% was achieved compared to pre- training on the ImageNet dataset or 12% when only children’s fashion was considered for training. / Modebranschen förväntas i framtiden vara online, och därför bedrivs det mycket forskning inom området bildanalys av modebilder. En uppgift som för närvarande innebär en stor arbetsbörda för nätbutiker är att manuellt tagga nya plagg med attribut som kategori, färg, mönster eller stil. Därför har omfattande forskning genomförts om automatisk förutsägelse av klädkategorier och attribut, och man har uppnått lovande resultat. Trots detta har ingen tidigare studie hittats i litteraturen som specifikt speglar prestandan för igenkänning av klädattribut för barnkläder. Syftet med det här arbetet är att fylla denna lucka och, som en del i en analys av mode, på ett effektivt sätt visa hur en modell som tränats för vuxenmode presterar jämfört med en modell som enbart tränats för barnmode. När man undersöker den globala förståelsen för barnkläder visar experimenten att den bästa prestandan uppnås när man utnyttjar domänkunskapen om vuxenmode, särskilt från iMaterialist- dataset, där man med den bästa modellen uppnådde en skillnad i den totala prestandan på cirka 3% jämfört med förträning på ImageNet- dataset eller 12% när endast barnmode beaktades vid träningen.