Japanese University Finds 4 Nazca Drawings with AI
![](https://japannews.yomiuri.co.jp/wp-content/uploads/2023/06/10580460.jpg)
One of the Nazca Lines geoglyphs newly identified using artificial intelligence, a drawing of a bird, is seen in Peru.
12:51 JST, June 3, 2023
YAMAGATA (Jiji Press) — Researchers at Yamagata University have identified four Nazca Lines geoglyphs in Peru using artificial intelligence deep learning technology in a joint project with IBM Japan Ltd.
According to the university, the use of AI helped significantly reduce the time needed to identify Nazca drawings. Previously, researchers looked over aerial photographs of wide areas to spot such geoglyphs.
The university’s Institute of Nazca has so far discovered 358 such drawings.
“We learned that work that could take more than 20 years for humans to do can be done in one year with AI,” Yamagata University Prof. Masato Sakai said.
“Research on Nazca geoglyphs is accelerating, and the discovery of more geoglyphs is expected to allow us to find out what they were drawn for,” Sakai added.
The institute developed a model to automatically detect new geoglyphs by getting AI to learn simplified patterns for divided parts of known geoglyphs, whose designs are unique and complicated.
The four Nazca drawings identified by AI are a pair of legs, a fish, a bird and a humanoid, measuring 78 meters, 19 meters, 17 meters and 5 meters, respectively.
The findings have been published in the Journal of Archaeological Science.
"Society" POPULAR ARTICLE
-
Japan Court OKs Sex Change without Surgery
-
3 Climbers Die On Mt. Fuji Within 2 Days Of Opening; Japan Police, Guides Urge Climbers To Prepare Well, Make Wise Decisions
-
Sex Crime Perpetrators Linked to U.S. Military in 166 Cases in Japan over 35 years; Local, Prefectural Governments Often Not Aware of Crimes
-
New Mt. Fuji Rules Reduce ‘Bullet Climbers’ by 90%; Access to Japan’s Iconic Peak Limited from Yamanashi Pref. Side
-
Tokaido Shinkansen Trains Suspended Between Hamamatsu and Nagoya Due to Accident; Resuming Services Expected Noon at Earliest
JN ACCESS RANKING