Large-Scale Fashion Database #DeepFashion#
We contribute DeepFashion database, a large-scale clothes database, which has over 800,000 diverse fashion images ranging from well-posed shop images to unconstrained consumer photos. DeepFashion is annotated with rich information of clothing items. Each image in this dataset is labeled with 50 categories, 1,000 descriptive attributes, bounding box and clothing landmarks. DeepFashion contains over 300,000 cross-pose/cross-domain image pairs.
The data can be downloaded from the
A Face Detection Benchmark #WIDER Face#
We conducted a large-scale face detection benchmark, containing 32,203 images and 393,703 face annotations, which are ten times larger than the exitsing datasets. More details
can be found in the
. The data can be downloaded from the
A Large-Scale Face Attributes Dataset #CelebA Dataset#
We presented a large-scale face attribute database, which contains 200K face images.
Each image was labeled with 40 facial attributes and five landmarks. More details
can be found in this
The Comprehensive Cars (CompCars) dataset contains data from two scenarios, including
images from web-nature and surveillance-nature. The web-nature data contains 163
car makes with 1,716 car models. There are a total of 136,726 images capturing the
entire cars and 27,618 images capturing the car parts. Please refer to our paper
for the details. Project Page
Pedestrian Attribute Recognition Database #PETA Database#
We release a new pedestrian attribute dataset, which is by far the largest and most
diverse of its kind. We present the benchmark performance by SVM-based method and
propose an alternative approach that exploits context of neighboring pedestrian
images for improved attribute inference.
The PETA dataset consists of 19,000 images, with resolution ranging from 17×39 to
169×365 pixels, covering more than 60 attributes. Project Page
Robust Facial Landmark Detection and Attribute Learning #ECCV 2014#
Facial landmark detection has long been impeded by the problems of occlusion and
pose variation. Instead of treating the detection task as a single and independent
problem, we investigate to optimize facial landmark detection together with heterogeneous
but subtly correlated tasks, e.g. head pose estimation and facial attribute inference.
Our new model detects 68 landmarks and achieves the state-of-the-art result on the
300-W benchmark dataset (mean error of 9.15% on the challenging IBUG subset). See
technical report for details.
Z. Zhang, P. Luo, C. C. Loy, X. Tang, Facial Landmark Detection by Deep Multi-task
Learning, ECCV 2014
PDF Technical Report
This Technical Report proposes a novel deep neural net, named multi-view perceptron
, which can untangle the identity and view features, and infer
a full spectrum of multi-view images in the meanwhile, given a single 2D face image.
The identity features of MVP achieve superior performance on the MultiPIE dataset.
MVP is also capable to interpolate and predict images under viewpoints that are
unobserved in the training data.
Zhenyao Zhu, Ping Luo, Xiaogang Wang, Xiaoou Tang, Deep Learning Multi-View Representation
for Face Recognition, Technical Report, arXiv:1406.6947, 2014