In [3] video of a speaker is used to train the distribution of visual parameters for each viseme, with synthesis performed by generating a trajectory that passes through the relevant distributions.
This article introduces the methods of reducing the utterance influences in visual parameters for the audio-visual-based emotion recognition.
The true and estimated visual parameters for the case of full covariance matrices with N = 28 states, M = 5 mixtures and co-articulation parameter t c = 3 are represented in Figure 8, where a good agreement between them can be observed.
However, most of existing work still simply combines the audio-visual parameters for emotion recognition with model of feature-level fusion or decision-level fusion (e.g., [19, 20, 45 47]), some of them just focus on getting Action Units (AUs) from facial expression rather than emotion recognition (e.g., [36 39, 41 44, 48, 49]).
Second, although the stimulus dimension modified in these experiments is brightness (emitted light), we have not attempted to exclude all other visual parameters as an origin for our observations.
L-dopa therapy in PD patients may also have an effect on visual parameters.
A joint AV-HMM was proposed to represent the audio-visual data and an algorithm for HMM inversion was derived for the estimation of the visual parameters, considering full covariance matrices for the observations.
For example, in active comparison tasks such visual confounds can have a profound effect on performance even when attempts are made to control for visual parameters across the whole experiment (Fuhs & McNeil, 2013; Gebuis & Reynvoet, 2011; Gebuis & Reynvoet, 2012a, 2012b; Gebuis & van der Smagt, 2011; Gilmore et al., 2013; Szűcs et al., 2013).
Finally, given a sequence of acoustic observations O a and the trained AV-HMM λ a v, the sequence of visual observations Õ v can be estimated by applying (7) to compute the visual parameters o ~ vt for each time t.
Comparison between the final and the initial visual parameters showed statistical similarity for both BDVA-grade (P = 0.96, Table 4) and logMAR vision score (P = 0.73, Figure 2), as is also indicated with a fairly even distribution above and beneath the line in Table 4 and in Figure 2. Posterior segment OGIs are associated with poor visual and anatomical outcomes [ 8– 12].
