论文标题
通过超级像素和图形神经网络无监督的图像语义分割
Unsupervised Image Semantic Segmentation through Superpixels and Graph Neural Networks
论文作者
论文摘要
在许多现实世界中，标记的数据稀缺性的许多实际情况下，无监督的图像分割是一项重要任务。在本文中，我们提出了一种新颖的方法，该方法利用相互信息最大化（MIM），神经超像素分段和图形神经网络（GNNS）的结合来利用无监督学习的最新进展，以端到端的方式，一种尚未探索的方法。我们利用超级像素的紧凑表示，并将其与GNN结合在一起，以学习图像的强大和语义上有意义的表示。具体而言，我们表明我们的基于GNN的方法允许在图像中的遥远像素之间建模相互作用，并在现有CNN之前充当强度，以提高准确性。我们的实验揭示了我们方法的定性和定量优势与四个流行数据集的当前最新方法相比。
Unsupervised image segmentation is an important task in many real-world scenarios where labelled data is of scarce availability. In this paper we propose a novel approach that harnesses recent advances in unsupervised learning using a combination of Mutual Information Maximization (MIM), Neural Superpixel Segmentation and Graph Neural Networks (GNNs) in an end-to-end manner, an approach that has not been explored yet. We take advantage of the compact representation of superpixels and combine it with GNNs in order to learn strong and semantically meaningful representations of images. Specifically, we show that our GNN based approach allows to model interactions between distant pixels in the image and serves as a strong prior to existing CNNs for an improved accuracy. Our experiments reveal both the qualitative and quantitative advantages of our approach compared to current state-of-the-art methods over four popular datasets.
