Image and video processing

ViStaCoRe: Visual Statistics Coding and Restoration Toolbox

The ViStaCoRe Coding Package is a Matlab Toolbox for achromatic and color image compression that includes a set of transform coding algorithms based on (1) Human Vision Models of different accuracy, and (2) coefficient selection through Sparse Regression in local frequency domains (in particular SVR). The ViStaCoRe Restoration Package is a Matlab Toolbox for image restoration that includes (1) classical regularization techniques, (2) classical wavelet thresholding techniques, (3) regularization functionals based on non-linear Human Vision models, and (4) denoising techniques based on Kernel regression in wavelet domains.

References
  • Image denoising with kernels based on natural image relations Laparra, V. and Gutiérrez, J. and Camps-Valls, G. and Malo, J. Journal of Machine Learning Research 11 :873-903, 2010
  • On the suitable domain for SVM training in image coding Camps-Valls, G. and Gutiérrez, J. and Gómez-Pérez, G. and Malo, J. Journal of Machine Learning Research 9 :49-66, 2008
  • Regularization operators for natural images based on nonlinear perception models Gutiérrez, J. and Ferri, F.J. and Malo, J. IEEE Transactions on Image Processing 15 (1):189-200, 2006
  • Nonlinear image representation for efficient perceptual coding Malo, J. and Epifanio, I. and Navarro, R. and Simoncelli, E.P. IEEE Transactions on Image Processing 15 (1):68-80, 2006
  • VistaQualityTools

    VistaQualityTools is a Matlab Toolbox for full reference color (and also achromatic) image quality assessment based on divisive normalization Human Vision models in the DCT and the Wavelet domains.

    References
  • Divisive normalization image quality metric revisited Laparra, V. and Muñoz-Marí, J. and Malo, J. Journal of the Optical Society of America A: Optics and Image Science, and Vision 27 (4):852-864, 2010
  • SpatioSpectralTools

    SpatioSpectralTools is a Matlab Toolbox for reflectance and illuminant estimation that uses spatial information to simplify the (otherwise ill-conditioned) inverse problem. The proposed analysis is useful to derive the spatio-spectral resolution (number of spectral sensors and spatial extent of ) required to solve a retrieval problem.

    References
  • The role of spatial information in disentangling the irradiance-reflectance-transmittance ambiguity Jimenez, S. and Malo, J. IEEE Transactions on Geoscience and Remote Sensing 52 (8):4881-4894, 2014
  • BasicVideoTools

    A Matlab Toolbox with convenient functions to handle video data. It includes routines to read VQEG and LIVE databases, generate synthetic sequences with controlled 2d and 3d speed (colored noise, random dots, rigid solid moving in a field, sequences from images), Spatio-temporal Fourier transforms, Perceptual sensors and filters (V1 and MT cells), and spatio-temporal CSFs.

    References
  • Importance of quantiser design compared to optimal multigrid motion estimation in video coding Malo, J. and Ferri, F.J. and Gutierrez, J. and Epifanio, I. Electronics Letters 36 (9):807-809, 2000
  • Video quality measures based on the standard spatial observer Andrew B. Watson and Jesus Malo ICIP (3), 2002
  • VideoQualityTools

    VideoQualityTools is a Matlab Toolbox for perceptual video quality assessment based on the Standard Spatial Observer model augmented with Divisive Normalization. It performed second-best in VQEG Phase-I using no ad-hoc hand-crafted features.

    References
  • Importance of quantiser design compared to optimal multigrid motion estimation in video coding Malo, J. and Ferri, F.J. and Gutierrez, J. and Epifanio, I. Electronics Letters 36 (9):807-809, 2000
  • Video quality measures based on the standard spatial observer Andrew B. Watson and Jesus Malo ICIP (3), 2002
  • VideoCodingTools

    VideoCodingTools is a Matlab Toolbox for motion estimation/compensation and video compression. Optical flow computation is done with perceptually meaningful hierarchical block matching, and residual quantization is done according to non-linear Human Visual System models.

    References
  • Importance of quantiser design compared to optimal multigrid motion estimation in video coding Malo, J. and Ferri, F.J. and Gutierrez, J. and Epifanio, I. Electronics Letters 36 (9):807-809, 2000
  • Video quality measures based on the standard spatial observer Andrew B. Watson and Jesus Malo ICIP (3), 2002