Extraction of Feature Quantities for Drawing a Good Portrait using a Gaze Point Measurement
注視点計測による似顔絵を上手に描くための特徴量抽出

Gen ONISHI, Toshiki KASHIO, Syu YORIFUJI, Ayaka KOCHI, Tsuyoshi SYOJI
2016 Transactions of Japan Society of Kansei Engineering  
This paper presents to extract examinees' feature quantities for a portrait when they draw it using an eye tracking system. We acquired the gaze coordinates of 40 healthy examinees' both eyes by non-restraint and noncontact measurement using infrared rays while they draw a portrait for 390 to 450 seconds. Eight art teachers in high school or university decided each 10 pictures of the high and low reputation value from their evaluation questionnaires. The results of analyzing the frequency of
more » ... glance translation a minute between the drawing canvas and the portrait model, the high reputation value group's frequency increased more than the low group's one. Moreover, the drawing glance rate (Ratio of the useful glance movement and the useless glance movement) proposed by the authors became a comparable results too. We succeeded in extracting the feature quantities to draw a portrait well as the numerical value, such as observing the model carefully and concentrating to write a portrait.
doi:10.5057/jjske.tjske-d-16-00035 fatcat:nfdjgdssr5en7doa6rp3gufvge