Fashion Landmark Detection in the Wild
This work addresses the need for more discriminative representations in fashion image understanding, offering a domain-specific advancement over previous methods using bounding boxes or human joints.
The paper tackles the problem of visual fashion analysis by introducing fashion landmark detection to predict functional key points on clothing items, and presents a dataset of over 120K images with eight landmarks per image, showing that this representation improves accuracy in applications like attribute prediction and clothes retrieval.
Visual fashion analysis has attracted many attentions in the recent years. Previous work represented clothing regions by either bounding boxes or human joints. This work presents fashion landmark detection or fashion alignment, which is to predict the positions of functional key points defined on the fashion items, such as the corners of neckline, hemline, and cuff. To encourage future studies, we introduce a fashion landmark dataset with over 120K images, where each image is labeled with eight landmarks. With this dataset, we study fashion alignment by cascading multiple convolutional neural networks in three stages. These stages gradually improve the accuracies of landmark predictions. Extensive experiments demonstrate the effectiveness of the proposed method, as well as its generalization ability to pose estimation. Fashion landmark is also compared to clothing bounding boxes and human joints in two applications, fashion attribute prediction and clothes retrieval, showing that fashion landmark is a more discriminative representation to understand fashion images.