Method for the production of sequences at required levels of interestingness
The invention is a method for automatically composing series of auditory events that unfold in time or visual events that unfold in space, which are statistically varied to be interesting. It also enables the creation of tonal series that are optimally interesting, given a provided set of tones. It allows also creating psychologically interesting versions of a given image.
The method computes a variable that operationalizes how interesting an event is in a given spatial or temporal context.
This invention introduces a method for producing sequences of events with specific levels of “interestingness”. It is applicable to auditory, visual, or audio-visual sequences, which can find utility in various environments, including video games or as backgrounds for online activities.
The method quantifies interestingness, distinguishing it from surprise. It can be employed to generate variations of existing compositions, dynamic visuals, or adjust the interestingness of images based on color palette revisions. An apparatus retrieves atomic elements from a database and generates auditory sequences with the specified interestingness level, or produces more/less interesting modifications of visual images, as required by the user. The method converts probability values into interestingness, simplifying the creation of sequences with defined levels of interest. This system is designed to adapt stimuli to user behavior, thereby enhancing engagement.
The technology is currently at TRL 3-4, as it has demonstrated its functionality in the domains of audio, dynamic video, and static images.
- Control VR or gaming auditory environments (background track) to maintain user interest.
- Dynamically control gaming dynamics and visuals to maximize engagement (not too surprising, not too little).
- Universal Applicability: Applicable to audio/visual content
- Human-Centric & Greater Ecological Valdity: takes into account human psychology
- Automated; no need for GUI.
- Adaptive Content Generation: interestingness can change based on measured behavior