Self-training solutions for the ICCV 2023 GeoNet Challenge
This work addresses domain adaptation for image recognition across geographical gaps, but it is incremental as it builds on existing source-free methods and benchmark challenges.
The paper tackled domain adaptation across geographical regions by proposing a two-stage source-free framework with a Swin Transformer backbone, achieving first place in the GeoUniDA challenge with an H-score of 74.56% and top-3 accuracies of 64.46% and 51.23% in other challenges.
GeoNet is a recently proposed domain adaptation benchmark consisting of three challenges (i.e., GeoUniDA, GeoImNet, and GeoPlaces). Each challenge contains images collected from the USA and Asia where there are huge geographical gaps. Our solution adopts a two-stage source-free domain adaptation framework with a Swin Transformer backbone to achieve knowledge transfer from the USA (source) domain to Asia (target) domain. In the first stage, we train a source model using labeled source data with a re-sampling strategy and two types of cross-entropy loss. In the second stage, we generate pseudo labels for unlabeled target data to fine-tune the model. Our method achieves an H-score of 74.56% and ultimately ranks 1st in the GeoUniDA challenge. In GeoImNet and GeoPlaces challenges, our solution also reaches a top-3 accuracy of 64.46% and 51.23%, respectively.