Nearest Neighbor Based Out-of-Distribution Detection in Remote Sensing Scene Classification
This addresses the challenge of reliable deployment in remote sensing for applications like geographic monitoring, but it is incremental as it adapts existing OOD detection methods to a new domain.
The paper tackles the problem of detecting out-of-distribution images in remote sensing scene classification, where deep learning models face inputs from different distributions than training data, and finds that a nearest neighbor-based method shows convincing advantages over maximum softmax probability.
Deep learning models for image classification are typically trained under the "closed-world" assumption with a predefined set of image classes. However, when the models are deployed they may be faced with input images not belonging to the classes encountered during training. This type of scenario is common in remote sensing image classification where images come from different geographic areas, sensors, and imaging conditions. In this paper we deal with the problem of detecting remote sensing images coming from a different distribution compared to the training data - out of distribution images. We propose a benchmark for out of distribution detection in remote sensing scene classification and evaluate detectors based on maximum softmax probability and nearest neighbors. The experimental results show convincing advantages of the method based on nearest neighbors.