Home

Awesome

SROSDA (ICCV 2021)

implementation of the ICCV work Towards Novel Target Discovery Through Open-Set Domain Adaptation [Paper].

:zap: Please check the extension journal work "Interpretable Novel Target Discovery Through Open-Set Domain Adaptation" (XSR-OSDA).

image

Data Preparation


DatasetDomainRole#Images#Attributes#Classes
D2AwAA <br> P <br> Rsource / target9,343 / 16,306 <br> 3,441 / 5,760 <br> 5,251 / 10,0478510 / 17
I2AwAI <br> Awsource / target2,970 / 37,3228540 / 50

(1) To extract pre-trained ResNet-50 features, check script:

./data/N2AwA/features/extract_resnet_features.ipynb

(2) Collect attributes for all samples based on their labels, check script:

./data/N2AwA/attributes/check_N2AwA_data.ipynb

Dependencies


Training


Step 1: Initialization clustering on target data (Seen/Unseen Initialization)

./data/N2AwA/refine_cluster-samples.ipynb

Note: Or use our clustering initialization results ./data/N2AwA/ directly.

Step 2: Train with the initialized clustering and pseudo labels on the extracted features.

python main.py

Evaluation


$OS^*$: class-wise average accuracy on the seen categories.

$OS^\diamond$: class-wise average accuracy on the unseen categories correctly classified as "unknown".

$OS$: $\frac{OS^* \times C_{shr} + OS^\diamond}{C_{shr} + 1}$

$C_{shr}$ is the number of shared categories between the source and target domains.

$S$: class-wise average accuracy on shared classes

$U$: class-wise average accuracy on unknown classes

$H = \frac{2 \times S \times U}{ S + U}$

Citation


If you think this work is interesting, please cite:

@InProceedings{Jing_2021_ICCV,
author = {Jing, Taotao and Liu, Hongfu and Ding, Zhengming},
title = {Towards Novel Target Discovery Through Open-Set Domain Adaptation},
booktitle = {International Conference on Computer Vision (ICCV)},
year = {2021}
}

Contact


If you have any questions about this work, feel free to contact