Feature Pyramid Network for Multi-class Land Segmentation

By Selim S. Seferbekov, Vladimir I. Iglovikov, Alexander V. Buslaev, Alexey A. Shvets. 2018

165 citations recorded by the source corpus.

Semantic segmentation is in-demand in satellite imagery processing. Because of the complex environment, automatic categorization and segmentation of land cover is a challenging problem. Solving it can help to overcome many obstacles in urban planning, environmental engineering or natural landscape monitoring. In this paper, we propose an approach for automatic multi-class land segmentation based on a fully convolutional neural network of feature pyramid network (FPN) family. This network is consisted of pre-trained on ImageNet Resnet50 encoder and neatly developed decoder. Based on validation results, leaderboard score and our own experience this network shows reliable results for the DEEPGLOBE - CVPR 2018 land cover classification sub-challenge. Moreover, this network moderately uses memory that allows using GTX 1080 or 1080 TI video cards to perform whole training and makes pretty fast predictions.

Subfields: Computer Vision and Pattern Recognition, Media Technology, Ocean Engineering

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Media Technology

In Ocean Engineering

Vladimir I. Iglovikov

Citations in the corpus

Albumentations: fast and flexible image augmentations (2018, methodology): “p Dih 4 [11]. Highperforming solutions to other challenges also implemented similar types of image augmentations, for example, for automatic road extraction [12] and for multi-class land segmentation [13] from satellite imagery. Fig. 3 provides examples of such image transforms applied to an image f”
Spatial Graph Convolutional Networks (2020, methodology): “The convolutional architectures were applied to numerous visual learning tasks on which they outperformed humans, e.g. image classiﬁcation [8], object detection [12] or image captioning [16].”
Camera Model Identification Using Convolutional Neural Networks (2018, background): “Problems in satellite [1], [2], [3], [4], medical [5], [6], [7], [8] or any other imagery are successfully tackled with these techniques, routinely beating human performance in many tasks, including classiﬁcation, segmentation, and detection.”
U-Net and Its Variants for Medical Image Segmentation: A Review of Theory and Applications (2020, background): “FPNs are designed to detect objects from each layer in the decoder and are particularly useful in producing multi-class segmentation maps [105].”
Imbalance Problems in Object Detection: A Review (2019, methodology): “Although FPN was originally proposed for object detection, it quickly became popular and has been used for different (but related) tasks such as shadow detection [107], instance segmentation [108], [109] and panoptic segmentation [110].”
Geometric Graph Convolutional Neural Networks (2019, background): “Convolutional Neural Networks (CNNs) outperform humans on visual learning tasks, such as image classiﬁcation (Krizhevsky, Sutskever, and Hinton 2012), object detection (Seferbekov et al. 2018) or image captioning (Yang et al. 2017).”