Tousch, Anne-Marie, Stéphane Herbin, and Jean-Yves Audibert. 2012. “Semantic Hierarchies for Image Annotation: A Survey.” Pattern Recognition 45 (1): 333–45. https://doi.org/10.1016/j.patcog.2011.05.017.

Notes

Semantic Level in Images:

1. The Of-ness vs. About-ness, i.e. the objective and concrete description vs. the subjective and abstract one. For instance, if ‘‘someone crying’’ is an objective description, ‘‘pain’’ is a subjective one.
2. The description can be generic, as in ‘‘a bridge’’, or specific, as in ‘‘Brooklyn bridge’’.
3. It can have four facets:
   1. time,
   2. localization,
   3. event,
   4. object

• Attributes in images

  • perceptual - objective
  • interpretive - subjective
  • reactive - “personal reactions of a person seeing the image” (Tousch et al., 2012, p. 335)
    

• levels of interpretation

  • image primitives
  • logical attributes
  • abstract attributes

In-text annotations