Image and Spectrum Based Deep Feature Analysis for Particle Matter Estimation with Weather Informatio

Two-Stream Non-Uniform Concentration Reasoning Network for Single Image Air Pollution Estimation

“…With the increasing availability of portable cameras and smart phones, directly estimating PM 2.5 based on digital photography shows advantages in efficiency and economic costs, and therefore has become a emerging research topic. Recently, the success of deep learning has yielded some neural network based approaches for estimating PM 2.5 concentration from images directly [2][3][4][5][6][7][8][9][10][11][12]. Chakma et al [5] classify natural images into three air pollution levels according to their PM 2.5 concentrations using VGG-19, and introduce CNN fine-tuning and CNN features-based random forest for transfer learning.…”