The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download
Publications Copernicus
Download
Citation
Articles | Volume XLIV-2/W1-2021
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLIV-2/W1-2021, 183–187, 2021
https://doi.org/10.5194/isprs-archives-XLIV-2-W1-2021-183-2021
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLIV-2/W1-2021, 183–187, 2021
https://doi.org/10.5194/isprs-archives-XLIV-2-W1-2021-183-2021

  15 Apr 2021

15 Apr 2021

HUMAN IMAGE MATTING BASED ON CONVOLUTIONAL NEURAL NETWORK AND PRINCIPAL CURVATURES

T. B. Sagindykov and E. A. Pavelyeva T. B. Sagindykov and E. A. Pavelyeva
  • Laboratory of Mathematical Methods of Image Processing, Faculty of Computational Mathematics and Cybernetics, Lomonosov Moscow State University, Moscow, Russia

Keywords: Image matting, Image segmentation, Principal Curvatures, Convolutional Neural Network, Deep learning

Abstract. Image matting often requires advanced image processing, especially in conditions, when small details such as hair are present in the image. In this article the hybrid method for human image matting based on convolutional neural network and principal curvatures is proposed. The U-Net based neural network is used to predict a rough foreground segmentation mask. Then the obtained foreground mask is refined by principal curvatures method to process the elongated hair-like structures. Test results show that the proposed method can improve the coarse human segmentation.