K-Covers for Active Learning in Image Classification

Deep learning has shown its effectiveness in various computer vision tasks. However, a large amount of labeled data is usually needed for deep learning approaches. Active learning can help reduce the labeling efforts by choosing the most informative samples to label and thus achieves a comparable performance with less labeled data. In this paper, we argue that only choosing samples based on some uncertainty function would lead to an unbalanced distribution of the selected samples, especially when the initial set of labeled samples are unbalanced. Following the intuition of reducing the repetitive sampling for similar images, we propose a novel K-Covers method to partition the feature space into several clusters and then choose one sample with the largest uncertainty in each cluster. Our method can constantly outperform the state-of-the-art with a clear margin.