煮不煮道

Abstract

中文摘要

大學生離家就學後，要面臨生活開銷與飲食健康問題。由於外食費用高昂，且營養成分難以掌握，若能培養自煮習慣，不僅能有效降低支出，也有助於建立健康的飲食型態。然而，多數大學生因缺乏烹飪經驗、飲食知識與菜單規劃能力，而難以長期實行自煮。因此，本專題嘗試結合人工智慧工具，設計一套能依使用者身體素質與現有食材，自動產生合適食譜的系統，以降低自煮門檻並提升實用性。

在研究過程中，我們運用 ChatGPT 與 Gemini 兩種生成式 AI 協助撰寫程式碼，並以 Google Colab 作為主要開發與執行平台。使用者可輸入身高、體重等身體數據，以及手邊現有的食材資訊，系統便能進行分析，產生符合需求的圖文食譜。此外，我們也加入圖片辨識功能，讓使用者上傳冰箱內食材照片，即可自動辨識蔬菜名稱，並提供相對應的料理方式與示意圖像，同時搭配清楚的步驟說明，引導烹飪新手完成料理。

結果，本系統成功整合飲食分析、食譜生成與圖片辨識等功能，能有效協助大學生進行自煮規劃。此成果不僅有助於提升飲食健康、減少剩食浪費，也能讓使用者在自煮過程中獲得生活成就感。 Abstract

After leaving home to attend university, college students have to face issues related to living expenses and healthy eating. Since eating out is costly and the nutritional content is hard to control, developing the habit of cooking at home can not only effectively reduce expenses but also help establish a healthy eating pattern. However, most college students find it difficult to maintain home cooking due to a lack of cooking experience, dietary knowledge, and menu planning skills. Therefore, this project attempts to integrate artificial intelligence tools to design a system that can automatically generate suitable recipes based on users' physical condition and available ingredients, aiming to lower the barrier to home cooking and enhance its practicality.

During the research process, we used the generative AI tools ChatGPT and Gemini to assist in writing code, and Google Colab served as the main development and execution platform. Users can input physical data such as height and weight, along with information on ingredients they have on hand, and the system will analyze the input and generate recipes that meet their needs. Additionally, we incorporated image recognition functionality, allowing users to upload photos of the ingredients in their refrigerator. The system can automatically identify the vegetable names, provide corresponding cooking methods and illustrative images, and offer clear step-by-step instructions to guide cooking beginners in completing the dishes.

As a result, this system successfully integrates dietary analysis, recipe generation, and image recognition functions, effectively assisting college students in planning their meals. This achievement not only helps improve diet health and reduce food waste but also allows users to gain a sense of accomplishment through the cooking process.