For the creation of the summaries, we have used the classifier's output, consisting of frames classified as salient; thus, segments or frames, chosen based on high confidence scores, form a binary indicator function curve, representing the most salient audio-visual-text events.
That is, we hypothesized that the link between facial movements and vocalizations' amplitude modulations [29], [30], [31] created a particularly salient audio-visual combination and that this induced anxiety.
Specifically, it is an improved synergistic approach to the problem of audio-visual salient event detection and movie summarization, also employing text saliency computation, initially presented in ICIP15 [14].
Discussion was directed by a focus group guide and vignette a digitally recorded audio story constructed from salient interview findings depicting a Black African or Black Caribbean woman's delayed presentation and diagnosis.
The additional annotation schemes, thus the audio-visual and cross-media events, and the emotion annotation can be used for audio-visual event classification, tracking, categorization of salient events regarding the above named schemes, or even for exploitation of the most relevant events in the automatically produced summaries (towards the end of creating user-defined summaries).
Two researchers trained in qualitative procedures conducted the interviews, which were audio-recorded, transcribed, and coded to identify salient issues.
In the following sections, we describe the results obtained with AFs on several difficult audio classification tasks, and report on the most salient aspects of these studies.
We observe that audiovisual salient segments correlate mostly with Token-Token, Metaphor, and Metonymy relations, audio salient segments correlate mostly with Token-Token, Metonymy, and Type-Token, visual salient segments correlate with Metaphor and Type-Token, while semantic salient segments with Token-Token and Metaphor.
Greater improvement can be seen for the monomodal salient event detection than the multimodal one and specifically for the audio modality (A-A evaluation).
A salient disadvantage to the PSN network could lie in the audio-recorded nature of the test events, as paralinguistic features indicating biases are excluded from the standardisation process.
