Aesthetic Quality Inference Engine (ACQUINE)
Mathematically formulate the core aesthetics and emotions prediction problems：
most dominant emotion
It is believed that great works of art evoke a mix of emotions leaving little space for emotion purity, clarity, or consistency. Learning a distribution of emotions from pictures requires a large and reliable emotion ground truth dataset.
Flickr’s interestingness attribute is another example of a community-driven measure of appeal based on user-judged content and community reinforcement.
Algorithm totally dependent on visual characteristics, image metadata
Artistic use of paint and brush can evoke a myriad of emotions among people, they are tools that artists employ to convey their ideas and feelings visually, semantically, or symbolically. Painting styles and brushstrokes are best understood and explained by art connoisseurs.
emotional congruence theory
For the unsupervised learning, feature similarity between data points is the driving factor for pattern discovery.
The difficulty of my theory of logical portrayal was that of finding a connection between the signs on paper and a situation outside in the world. I always said that truth is a relation between the proposition and the situation, but could never pick out such a relation (19e-20e; quoted in Word and World, 71). -Wittgenstein ‘picture theory of language’