基於3D食物資料庫之擺盤設計

Abstract

隨著現今社會中對飲食文化的日益重視,除了食物本身滋味,食物擺盤設計也成為一 項重要指標。如諺語“You eat with your eyes first.”所言,研究已證實,食物擺盤藝 術除了具美觀之用,也影響著食用者的心情及食欲,此現象對於幼童尤其顯著。而食物擺 盤通常是由數量不定的數種食材所構成,利用食材間的堆疊來表現其目標圖案的輪廓及顏 色,其中變化之廣使其成為研究議題。然而,無法決定食材的選擇以及其間的擺放方式, 或是設計擺盤過程中必經的反覆嘗試擺放,都可能被視為嘗試擺盤設計的困難點而令人卻 步。因此,為了提高擺盤藝術之自由度及方便性,讓更多人能夠體會擺盤藝術,我們提出 了一個新的方法來引導使用者設計食物擺盤,使用者只需要輸入其欲擺放的目標圖案,我 們可藉由我們的形狀配對演算法從食物資料庫中找出每個食材可能放置的位置,在這數組 可能位置中,自動地透過形狀、顏色及邊緣的符合程度定義相似度,在短時間內分別選出 最好的食材元件及其擺放位置,直至所有目標圖案邊緣皆被滿足,最後輸出堆疊結果及各 食材對應之旋轉方式,因不容許形變及大小比例之變化,我們能夠得到自然而能夠在現實 世界完成的擺盤設計。

Plate presentation plays an important role nowadays. Just like the proverb says, “you eat with your eyes first”. People expect foods to be not only delicious but also visually pleasing. Several researches also have shown that a good plate presentation can attract attention from people and lead them into a good mood. The influence is significant particularly to children. In general, a plate presentation is the combination of foods that appears like a cartoon character or an interesting pattern. Therefore, the goal is to arrange foods at different positions and layers to approximate colors and boundaries of a pattern. Given that the task is difficult to novice users who have bad sense of shape analysis and are not talent to aesthetics, even a bad design would cost them a lot of time and efforts. Accordingly, we introduce a fully automatic algorithm to generate plate presentations. The presented system analyzes the input pattern and determines an optimal arrangement of foods to approximate the pattern's colors and contours. We claim that the results are physically achievable in the real world because scales and shapes of foods are retained by our system. Compared to a manual design that is time consuming, our system computes a comparable plate presentation within five minutes, as demonstrated in our results.