Deep Curriculum Learning for PolSAR Image Classification

TL;DR

This paper introduces a deep curriculum learning approach for PolSAR image classification that uses entropy-alpha decomposition to assess patch complexity and employs a pacing function to improve accuracy and training speed.

Contribution

It presents a novel curriculum learning method specifically designed for PolSAR data, integrating complexity estimation and a pacing function for enhanced performance.

Findings

Improved classification accuracy on AIRSAR Flevoland dataset

Faster training convergence with the proposed method

Effective complexity assessment of PolSAR patches

Abstract

Following the great success of curriculum learning in the area of machine learning, a novel deep curriculum learning method proposed in this paper, entitled DCL, particularly for the classification of fully polarimetric synthetic aperture radar (PolSAR) data. This method utilizes the entropy-alpha target decomposition method to estimate the degree of complexity of each PolSAR image patch before applying it to the convolutional neural network (CNN). Also, an accumulative mini-batch pacing function is used to introduce more difficult patches to CNN.Experiments on the widely used data set of AIRSAR Flevoland reveal that the proposed curriculum learning method can not only increase classification accuracy but also lead to faster training convergence.