Draft #25
Facial recognition science is a very interesting and is a field relatively little known by the public. Firstly, research of facial recognition by humans has been performed, and the question asked was "What kind of biases are there that determine ability to recognize faces and identify their characteristics?". Susan Mason examined how age corresponded to ability to identify faces and recall names connected with the faces based on age similarity between the subject and the example. She found that people were more likely to recall the names and recognize faces as familiar of people who were similar in appearance to themselves, namely age being a factor in facial similarity between subject and example. This means that older people recognized and recalled the names of other elderly people in the experiment more often than they did younger people, and vice versa for younger people. (Mason)
Now this is all relative to the human ability, but with modern technology we no longer have to rely on the human capacity. Software has been developed to interpret faces and characteristics, identify key features, and assign a name and gender to a face. There are draw backs to this unperfected science, however. University of Colorado Boulder's Lisa Marshall published an article examining the ability and inability of this technology. Many different softwares are coming out that are able to accurately identify age and gender, amoung cisgendered people. That is, born females and males that identify with their gender are categorically matched to their genders more than 95% of the time. However, when the aoftware attempts a gender identification for trans and nonbinary gendered persons, the data does not look as good. This idea of rendering the gender identify of an individual by a photo completely falls apart when their characteristics and identity do not align for the software. (Marshall)
Susan E. (1986) Age and gender as factors in facial recognition and identification, Experimental Aging Research, 12:3, 151-154, DOI: 10.1080/03610738608259453
Marshall, Lisa. “Facial Recognition Software Has a Gender Problem.” CU Boulder Today, University of Colorado Boulder, 9 Oct. 2019, www.colorado.edu/today/2019/10/08/facial-recognition-software-has-gender....
Recent comments