Few-shot semantic segmentation via mask aggregation
This addresses the problem of segmenting novel classes with limited labeled data for computer vision researchers, but it is incremental as it offers an alternative baseline method rather than a major breakthrough.
The paper tackles few-shot semantic segmentation by proposing a mask aggregation network (MANet) that generates masks and their probabilities instead of using pixel-wise classification, achieving competitive performance with state-of-the-art methods on PASCAL-5^i and COCO-20^i datasets.
Few-shot semantic segmentation aims to recognize novel classes with only very few labelled data. This challenging task requires mining of the relevant relationships between the query image and the support images. Previous works have typically regarded it as a pixel-wise classification problem. Therefore, various models have been designed to explore the correlation of pixels between the query image and the support images. However, they focus only on pixel-wise correspondence and ignore the overall correlation of objects. In this paper, we introduce a mask-based classification method for addressing this problem. The mask aggregation network (MANet), which is a simple mask classification model, is proposed to simultaneously generate a fixed number of masks and their probabilities of being targets. Then, the final segmentation result is obtained by aggregating all the masks according to their locations. Experiments on both the PASCAL-5^i and COCO-20^i datasets show that our method performs comparably to the state-of-the-art pixel-based methods. This competitive performance demonstrates the potential of mask classification as an alternative baseline method in few-shot semantic segmentation. Our source code will be made available at https://github.com/TinyAway/MANet.