Published on Sun Apr 19 2020

An end-to-end CNN framework for polarimetric vision tasks based on polarization-parameter-constructing network

Yong Wang, Qi Liu, Hongyu Zu, Xiao Liu, Ruichao Xie, Feng Wang

Polarization information cannot be fully utilized in convolutional neural network(CNN) A novel end-to-end CNN framework for polarization vision tasks is proposed. The framework consists of two sub-networks: a polarimetric-parameter-constructing network and a task network.

0
0
0
Abstract

Pixel-wise operations between polarimetric images are important for processing polarization information. For the lack of such operations, the polarization information cannot be fully utilized in convolutional neural network(CNN). In this paper, a novel end-to-end CNN framework for polarization vision tasks is proposed, which enables the networks to take full advantage of polarimetric images. The framework consists of two sub-networks: a polarization-parameter-constructing network (PPCN) and a task network. PPCN implements pixel-wise operations between images in the CNN form with 1x1 convolution kernels. It takes raw polarimetric images as input, and outputs polarization-parametric images to task network so as to complete a vison task. By training together, the PPCN can learn to provide the most suitable polarization-parametric images for the task network and the dataset. Taking faster R-CNN as task network, the experimental results show that compared with existing methods, the proposed framework achieves much higher mean-average-precision (mAP) in object detection task

Wed Oct 02 2019
Machine Learning
Road scenes analysis in adverse weather conditions by polarization-encoded images and adapted deep learning
Object detection in road scenes is necessary to develop both autonomous vehicles and driving assistance systems. The efficiency of the proposed method is mostly due to the high power of the polarimetry to discriminating any object by its reflective properties.
0
0
0
Fri May 22 2020
Computer Vision
Polarimetric image augmentation
Polarization images are not only scalar (intensity) images and standard augmentation techniques cannot be applied straightforwardly. We propose to enhance deep learning models through a regularized augmentation procedure. We subsequently observe an average of 18.1%improvement in IoU.
0
0
0
Thu Jan 09 2020
Computer Vision
Objects detection for remote sensing images based on polar coordinates
0
0
0
Thu Oct 31 2019
Machine Learning
Very high resolution Airborne PolSAR Image Classification using Convolutional Neural Networks
We exploit convolutional neural networks (CNNs) for the classification of very high resolution (VHR) polarimetric SAR (PolSAR) data. We use the SegNet model for semantic segmentation, which corresponds to pixelwise classification in remote sensing.
0
0
0
Tue Nov 19 2019
Machine Learning
Eliminating artefacts in Polarimetric Images using Deep Learning
Artefacts can range from internal reflections in a telescope to satellite trails that could contaminate an area of interest in the image. Polarization measurements done using Imaging Polarimeters are very sensitive to the presence of artefacts.
0
0
0
Mon Jul 09 2018
Computer Vision
Polarimetric Convolutional Network for PolSAR Image Classification
Polarimetric scattering matrix of synthetic aperture radar (PolSAR) data has always been the focus of PolSAR image classification. Based on sparse scattering coding and convolution neural network, the polarimetric convolutional network is proposed to classify PolSar images.
0
0
0