Module code >> skvideo.measure.brisque
Fork me on GitHub

Source code for skvideo.measure.brisque

import numpy as np
import scipy.fftpack
import scipy.io
import scipy.misc
import scipy.ndimage
import scipy.stats

from ..utils import *


def _extract_subband_feats(mscncoefs):
    # alpha_m,  = extract_ggd_features(mscncoefs)
    alpha_m, sigma_sq = ggd_features(mscncoefs.copy())
    pps1, pps2, pps3, pps4 = paired_product(mscncoefs)
    alpha1, N1, bl1, br1, lsq1, rsq1 = aggd_features(pps1)
    alpha2, N2, bl2, br2, lsq2, rsq2 = aggd_features(pps2)
    alpha3, N3, bl3, br3, lsq3, rsq3 = aggd_features(pps3)
    alpha4, N4, bl4, br4, lsq4, rsq4 = aggd_features(pps4)
    return np.array([
            alpha_m, sigma_sq,
            alpha1, N1, lsq1**2, rsq1**2,  # (V)
            alpha2, N2, lsq2**2, rsq2**2,  # (H)
            alpha3, N3, lsq3**2, rsq3**2,  # (D1)
            alpha4, N4, lsq4**2, rsq4**2,  # (D2)
    ])


[docs]def brisque_features(videoData): """Computes Blind/Referenceless Image Spatial QUality Evaluator (BRISQUE) features. [#f1]_ [#f2]_ Since this is a referenceless image quality algorithm, only 1 video is needed. This function provides the raw features extracted per frame. Parameters ---------- videoData : ndarray Reference video, ndarray of dimension (T, M, N, C), (T, M, N), (M, N, C), or (M, N), where T is the number of frames, M is the height, N is width, and C is number of channels. Returns ------- features : ndarray A matrix of shape (T, 36) of the computed BRISQUE features. References ---------- .. [#f1] A. Mittal, A. K. Moorthy and A. C. Bovik, "No-Reference Image Quality Assessment in the Spatial Domain" IEEE Transactions on Image Processing, 2012. .. [#f2] A. Mittal, A. K. Moorthy and A. C. Bovik, "Referenceless Image Spatial Quality Evaluation Engine," 45th Asilomar Conference on Signals, Systems and Computers , November 2011. """ videoData = vshape(videoData) T, M, N, C = videoData.shape assert C == 1, "brisque_features called with video having %d channels. Please supply only the luminance channel." % (C,) feats = np.zeros((T, 36), dtype=np.float32) for i in range(T): full_scale = videoData[i, :, :, 0].astype(np.float32) half_scale = scipy.misc.imresize(full_scale, 0.5, interp='bicubic', mode='F') full_scale, _, _ = compute_image_mscn_transform(full_scale) half_scale, _, _ = compute_image_mscn_transform(half_scale) feats[i, 18*0:18*1] = _extract_subband_feats(full_scale) feats[i, 18*1:18*2] = _extract_subband_feats(half_scale) return feats