Fusion of Heterogeneous Data in Convolutional Networks for Urban Semantic Labeling (Invited Paper)

01/20/2017

∙

In this work, we present a novel module to perform fusion of heterogeneous data using fully convolutional networks for semantic labeling. We introduce residual correction as a way to learn how to fuse predictions coming out of a dual stream architecture. Especially, we perform fusion of DSM and IRRG optical data on the ISPRS Vaihingen dataset over a urban area and obtain new state-of-the-art results.

READ FULL TEXT

Fusion of Heterogeneous Data in Convolutional Networks for Urban Semantic Labeling (Invited Paper)

Semantic Segmentation of Earth Observation Data Using Multimodal and Multi-scale Deep Networks

Beyond RGB: Very High Resolution Urban Remote Sensing With Multimodal Deep Networks

Multiresolution Fully Convolutional Networks to detect Clouds and Snow through Optical Satellite Images

Fully Convolutional Networks for Text Classification

Heterogeneous Data Fusion Considering Spatial Correlations using Graph Convolutional Networks and its Application in Air Quality Prediction

Classification of Hyperspectral and LiDAR Data Using Coupled CNNs

Fast Training of Convolutional Networks through FFTs

Fusion of Heterogeneous Data in Convolutional Networks for Urban Semantic Labeling (Invited Paper)

Related Research

Semantic Segmentation of Earth Observation Data Using Multimodal and Multi-scale Deep Networks

Beyond RGB: Very High Resolution Urban Remote Sensing With Multimodal Deep Networks

Multiresolution Fully Convolutional Networks to detect Clouds and Snow through Optical Satellite Images

Fully Convolutional Networks for Text Classification

Heterogeneous Data Fusion Considering Spatial Correlations using Graph Convolutional Networks and its Application in Air Quality Prediction

Classification of Hyperspectral and LiDAR Data Using Coupled CNNs

Fast Training of Convolutional Networks through FFTs