Avec 2014

Abstract

Mood disorders are inherently related to emotion. In particular, the behaviour of people suffering from mood disorders such as unipolar depression shows a strong temporal correlation with the affective dimensions valence, arousal and dominance. In addition to structured self-report questionnaires, psychologists and psychiatrists use in their evaluation of a patient's level of depression the observation of facial expressions and vocal cues. It is in this context that we present the fourth Audio-Visual Emotion recognition Challenge (AVEC 2014). This edition of the challenge uses a subset of the tasks used in a previous challenge, allowing for more focussed studies. In addition, labels for a third dimension (Dominance) have been added and the number of annotators per clip has been increased to a minimum of three, with most clips annotated by 5. The challenge has two goals logically organised as sub-challenges: the first is to predict the continuous values of the affective dimensions valence, arousal and dominance at each moment in time. The second is to predict the value of a single self-reported severity of depression indicator for each recording in the dataset. This paper presents the challenge guidelines, the common data used, and the performance of the baseline system on the two tasks.

Modeling Multiple Time Series Annotations as Noisy Distortions of the Ground Truth: An Expectation-Maximization Approach

“…Several previous works [8], [9], [46] have used this assumption in obtaining the ground truth from multiple annotators on similar time series modeling problems. This scheme assigns equal weight to each annotator and does not account for individual differences.…”

“…A classic example is modeling emotional dimensions (e.g. valence, dominance, arousal) during human interaction [9], human-computer interaction [8], [25] as well as in music [26], [27]. These studies use multiple annotators to derive the ground truth reference and use heuristic metrics over the annotator ratings as a proxy for the latent emotional dimension.…”

Quantitative comparison of motion history image variants for video-based depression assessment

“…The majority of published approaches materialized within the AVEC "Depression Recognition Sub-challenge" (DSC) [9][10][11].…”

“…The dataset consisted of video feeds of undiagnosed volunteers performing the following tasks: vowel pronunciation, solving a task out loud, counting from 1 to 10, reading novel excerpts, singing, and describing a specific scene displayed in pictorial form. During AVEC'13, the complete recording was used for testing the performance of the different approaches, while in AVEC'14, two tasks were selected: the Northwind (reading excerpt) and the Freeform (answering to questions) [10]. AVEC'16 [11] utilized a different dataset (DAIC-WOZ [12]) and provided only the features extracted with the OpenFace software [13], instead of the original video recordings.…”

“…They used the Local Gabor Binary Patterns in Three Orthogonal Planes, provided by the challenge organizers [10], achieving 82% accuracy. Similar results were obtained in the crosscultural study of Alghowinem et al [15] who focused on geometrical features derived from eye activity to achieve 81.3% classification accuracy utilizing one subset of the AVEC'13 dataset, among other datasets.…”