|
转自：爱可可-爱生活 论文《Fashion Landmark Detection in the Wild》摘要： Visual fashion analysis has attracted many attentions in the recent years.Previous work represented clothing regions by either bounding boxes or humanjoints. This work presents fashion landmark detection or fashion alignment,which is to predict the positions of functional key points defined on thefashion items, such as the corners of neckline, hemline, and cuff. To encouragefuture studies, we introduce a fashion landmark dataset with over 120K images,where each image is labeled with eight landmarks. With this dataset, we studyfashion alignment by cascading multiple convolutional neural networks in threestages. These stages gradually improve the accuracies of landmark predictions.Extensive experiments demonstrate the effectiveness of the proposed method, aswell as its generalization ability to pose estimation. Fashion landmark is alsocompared to clothing bounding boxes and human joints in two applications,fashion attribute prediction and clothes retrieval, showing that fashionlandmark is a more discriminative representation to understand fashion images. |
|
|
