The predictions model gives you detailed control over what is calculated in an analysis and how results are presented.

These predictions are currently supported:

name	Description	default configuration
attention	Which elements of the image draw the attention of the average person viewing it?	`{"outputs": ["attentionMap"]}`
clarity	As how clear or cluttered is the image perceived?	`{"outputs": ["score"]}`
excitingness	As how exciting is the image perceived?	`{"outputs": ["score"]}`

Attention outputs

name	Description
attentionMap	An attention heat-map showing a detailed distribution of the predicted fixation densities.
perceptionMap	Visualises which areas are immediately attract the attention of a viewer.
hotspotsMap	Uses circles to visualise the most attention-grabbing spots. The larger a circle, the more attention-grabbing the spot is.

Clarity outputs

output	Description
score	A score between 0 and 100 quantifying the clarity rating of the image.
map	A visualisation highlighting the degree to which individual areas of the image contribute to the overall perception of clarity or clutter.

🚧
Please note that due to technical reason, currently all available outputs are returned, regardless of which outputs are requested. This may change to returning only the requested outputs without notice.

Excitingness outputs

output	Description
score	A score between 0 and 100 quantifying the predicted excitingness rating of the image.

Examples:

An attention prediction using its default configuration:

{
  "attention": {}
}

Multiple predictions for the same input: a clarity prediction with both score and map outputs, and an excitingness prediction with its default configuration:

{ 
  "clarity": {
    "outputs": ["score", "map"]
  },
  "excitingness": {}
}