The code repository for "Learning Placeholders for Open-Set Recognition " [paper] (CVPR21) in PyTorch. If you use any content of this repo for your work, please cite the following bib entry:

@inproceedings{zhou2021learning,
author = {Zhou, Da-Wei and Ye, Han-Jia and Zhan, De-Chuan},
title = {Learning Placeholders for Open-Set Recognition},
booktitle = {CVPR},
pages = {4401-4410},
year = {2021}
}

Traditional classifiers are deployed under closed-set setting, with both training and test classes belong to the same set. However, real-world applications probably face the input of unknown categories, and the model will recognize them as known ones. Under such circumstances, open-set recognition is proposed to maintain classification performance on known classes and reject unknowns. The closed-set models make overconfident predictions over familiar known class instances, so that calibration and thresholding across categories become essential issues when extending to an open-set environment. To this end, we proposed to learn PlaceholdeRs for Open-SEt Recognition (Proser), which prepares for the unknown classes by allocating placeholders for both data and classifier. In detail, learning data placeholders tries to anticipate open-set class data, thus transforms closed-set training into open-set training. Besides, to learn the invariant information between target and non-target classes, we reserve classifier placeholders as the class-specific boundary between known and unknown. The proposed Proser efficiently generates novel class by manifold mixup, and adaptively sets the value of reserved open-set classifier during training. Experiments on various datasets validate the effectiveness of our proposed method.

The following packages are required to run the scripts:

• PyTorch-1.4 and torchvision

• Dataset: please download the dataset and unzip it into the folder data/cifar

• Pre-trained weights: we provide our pretrained weights for this trial. Download and put it to 'results/Dcifar10_relabel-Msoftmax-BWideResnet/LR0.1-K[8 4 7 2 1 9]-U[3 0 6 5]-Seed9/' You can also use the pretrain code for other tasks. See below for details.

Download CIFAR10 matlab version and unzip it to data/cifar/

There are four parts in the code.

• models: It contains the backbone network for the experiment
• data: Images and splits for the data sets, as well as the data preprocessing code.
• results: The pre-trained weights of different networks.

We provide the code to reproduce results on CIFAR 10 unknown detection task, c.f. Table.1 in the main paper.

• Download the pretrained weights:

• Train the Proser model

  python proser_unknown_detection.py
  

  and you will get corresponding AUC measures.

We thank the following repos providing helpful components/functions in our work.

• pytorch-cifar

• FEAT

If there are any questions, please feel free to contact with the author: Da-Wei Zhou (zhoudw@lamda.nju.edu.cn). Enjoy the code.