The goal of this work is to develop a novel learning framework for accurate and expressive fashion captioning.
Color compatibility is important for evaluating the compatibility of a fashion outfit, yet it was neglected in previous studies.
In this paper, we present new data pre-processing and augmentation techniques for DNN-based raw image denoising.
Surface-based geodesic topology provides strong cues for object semantic analysis and geometric modeling.
Being inspired by child's learning experience - taught first and followed by observation and questioning, we investigate a critically supervised learning methodology for object detection in this work.
Recent development in fully convolutional neural network enables efficient end-to-end learning of semantic segmentation.
Ranked #53 on Semantic Segmentation on NYU Depth v2