Convolutional Neural Networks for Fashion Classification and Object Detection

Fashion classification encompasses the identification of clothing items in an image. The field has applications in social media, e-commerce, and criminal law. In our work, we focus on four tasks within the fashion classification umbrella: (1) multiclass classification of clothing type; (2) clothing attribute classification; (3) clothing retrieval of nearest neighbors; and (4) clothing object detection. We report accuracy measurements for clothing style classification (50.2%) and clothing attribute classification (74.5%) that outperform baselines in the literature for the associated datasets. We additionally report promising qualitative results for our clothing retrieval and clothing object detection tasks.