Creative Emotions

Creative Emotions results from extensive research on relationships between Emotions, Image, and Music characteristics.


Through an open communication of these findings via a website and API (, we hope that other researchers, designers, or curious minds can easily access the results of these auditory-visual associations research and create emotion-sensitive design artifacts.


Figure 1

Website overview. Example of emotions associated to the visual property/value of “complexity > simple”.


The developed webpage contains all the associations present in this cross-modal dataset. These associations can be navigated from a “top down” or “bottom up” approaches (see “Creative Emotions” page). “Top-down” search begins with the properties that compose a piece of music or image, followed by the values of each property and its associated emotion. “Bottom-up” search begins with a list of emotions common to music and image domains, followed by the values associated with each emotion of the list. Finally, all the properties considered perceptually relevant in representing emotions are drawn with a cross “+”.


In addition, we provide a JSON file (see more on the “Documentation” page) containing this dataset, ready to be used in upcoming projects. The structure inside the JSON file can be accessed by typing the “ed” variable in the browser console. The information is organised like this: ed > domain > levels > properties > emotions. For example, there is a level called Harmony Melody in the Music domain. This level has a property named Sonance, which contains a value called Harmony Consonant. Furthermore, this value is associated with several Emotions, such as Happiness (e.g. Music > Harmony Melody > Harmony Consoant > Happiness.).


The JSON file also allows us to access the level of information desired directly. For example, we can access the property and see all associated emotions or access emotion and see all associated properties. Generally speaking, we can search the information tree up (bottom-up) and down (top-down).


Examples of projects performed with the creative emotions dataset can be consulted on the “Demonstration” page. We hereby preview two projects developed with the Creative Emotions dataset: “Face Emotions” and “Found in Translation”.


Face Emotions“ is a web application exploring the potential of integrating an interdisciplinary dataset (creative emotions) to create computational design artifacts.


Figure 2

Face Emotions App: Angry, Calm, Happy, Sad visual representations. Developed by Joana Oliveira.


Found in translation“ is part of a research project where we propose a novel approach to the computational exploration of relationships between music and abstract images, grounded by findings from cognitive sciences (emotion and perception). For more information on this, please consult the following research article: Rodrigues, A., Sousa, B., Cardoso, A., & Machado, P. (2022). “Found in Translation”: An Evolutionary Framework for Auditory–Visual Relationships. Entropy24(12), 1706.


Figure 3

Example of image generation based on creative emotions dataset (Rodrigues et al. 2022).


For new contents regarding this dataset or to add a new project, please e-mail us: anatr(at)


Joana Oliveira

Ana Rodrigues

Daniel Lopes

Amílcar Cardoso

Penousal Machado