Identifying People in Old Photographs Using Facial Recognition Technology - Application to “Verbeck Photograph”

The names of people in old photographs are often incorrect as a result of later generations' incorrect guesses or the application of convenient people. The author has therefore applied AI-based facial recognition technology as an objective means of identifying the person in the “Verbeck Photogr...

Full description

Saved in:
Bibliographic Details
Published inJournal of The Society of Photographic Science and Technology of Japan Vol. 87; no. 2; pp. 148 - 154
Main Author Ito, Yasuo
Format Journal Article
LanguageJapanese
Published THE SOCIETY OF PHOTOGRAPHY AND IMAGING OF JAPAN 20.05.2024
社団法人 日本写真学会
Subjects
Online AccessGet full text
ISSN0369-5662
1884-5932
DOI10.11454/photogrst.87.148

Cover

More Information
Summary:The names of people in old photographs are often incorrect as a result of later generations' incorrect guesses or the application of convenient people. The author has therefore applied AI-based facial recognition technology as an objective means of identifying the person in the “Verbeck Photograph”, which has long been the subject of much debate. Amazon Web Services' Amazon Rekognition R was used for face recognition to examine whether any of the people in the “Verbeck Photograph” corresponded to the Iwakura brothers, Tomosada and Tomotsune, who are considered most likely to be in the photo visually, as well as some samurai from the Saga clan. As a result, the author was able to identify 10 people from the Saga clan in the “Verbeck Photo” in addition to Iwakura brothers. On the other hand, it was also clear that Saigo Takamori, Sakamoto Ryoma, and other patriots during the end of Edo period were not identified in the photo, which is a popular myth. Face recognition technology is useful for the study of old photographs in order to evaluate them objectively without being bound by subjectivity.
ISSN:0369-5662
1884-5932
DOI:10.11454/photogrst.87.148