본문 바로가기 주메뉴 바로가기
22대 대한민국 국회 국회정보길라잡이 국회의원검색
회원가입 로그인
HOME

정보검색

소장정보 검색

국회도서관 홈으로 정보검색 소장정보 검색
결과 내 검색 동의어 포함
결과 내 검색 동의어 포함

-

목차보기

Title Page

Abstract

Contents

Ⅰ. INTRODUCTION 13

Ⅱ. RESEARCH HISTORY 15

2.1. The etymology and history of perfume 15

2.2. The definition and function of fragrance 17

2.3. The structure of perfume 18

2.3.1. The stage of perfume 18

2.3.1. The volatility of perfume 18

2.4. The role of perfume 19

2.5. Fragrance 20

2.5.1. Natural fragrance 20

2.5.2. Extraction methods of natural fragrances 20

SECTION 1. Development of an Automatic Perfume Production Machine 21

Ⅲ. MATERIALS AND METHODS 22

3.1. Auto perfume device 22

3.1.1. Auto perfume machine design concept 22

3.1.2. The design and production of auto perfume device component 32

3.1.3. Principle of pump operation of Automatic perfume device 35

3.1.4. Automatic perfume block diagram 37

3.1.5. Automatic perfume device's program design 39

Ⅳ. RESULTS AND DISCUSSION 41

4.1. Manufacturing of Automatic perfume device 41

4.1.1. Exterior structure of automatic perfume device 41

4.1.2. Implementation of automatic perfume device's program 44

SECTION 2. Artificial Intelligence-assisted Recommendation Program 55

Ⅲ. MATERIALS AND METHODS 56

3.1. Fragrance recipe recommendation system 56

3.1.1. Open data 56

3.1.2. Closed data 59

3.1. Plant materials 69

3.2. Essential oil extraction 69

3.3. Gas Chromatography-mass spectrometry (GC-MS) 70

Ⅳ. RESULTS AND DISCUSSION 71

4.1. Recipe recommendation system 71

4.1.1. Open data collection 71

4.1.2. Major fragrance type 72

4.2. Close data collection 74

4.2.1. The production of fragrances combination 74

4.2.2. Fragrance Evaluation program 77

4.2.3. Artificial intelligence algorithm evaluation 81

4.2.4. Implementation of perfume recipe recommendation system 82

4.2.5. Execution of Recipe Recommendation System Using Artificial Intelligence 91

4.3. The Yield and Color of Cymbopogon 104

4.4. Chemical composition of lemongrass essential oil from Cambodia 105

4.5. Discussion 107

Ⅴ. CONCLUSION 110

REFERENCES 114

논문요지 118

APPENDIX 119

List of Tables

Table 1. Pump speed of an automatic perfume device. 54

Table 2. Close data fragrance combination 61

Table 3. Fragrance type and characteristics of 1,026 perfume samples. 63

Table 4. Collected 54000 perfumes data for the construction of big data from the World Wide Web. 71

Table 5. A total of five fragrances combination. 76

Table 6. Excel data generated from AI program. 80

Table 7. 3 Algorithm training results. 82

Table 8. Citrus-floral-sweet perfume recommendation and intensity. 93

Table 9. Citrus-fruity-fresh perfume recommendation and intensity. 95

Table 10. Citrus-sweet-oriental perfume recommendation and intensity. 97

Table 11. Sweet-citrus-musky perfume recommendation and intensity. 99

Table 12. Citrus-acidic-sweet perfume recommendation and intensity. 101

Table 13. Aldehyde-minty-sweet perfume recommendation and intensity. 103

Table 14. Lemongrass essential oil yield. 104

Table 15. Chemical composition of lemongrass essential oil. 105

Table 16. . Chemical composition of lemongrass essential oil. 106

Table 17. Chemical composition of lemongrass essential oil. 106

List of Figures

Figure 1. The overall design of the perfume device. 24

Figure 2. The overall design of the perfume device's isometric view. 25

Figure 3. Raspberry Pi's frame design from the front. 26

Figure 4. The design of Raspberry Pi's frame from the bottom. 27

Figure 5. The design of the power supply panel. 28

Figure 6. The design of the relay board panel. 29

Figure 7. The design of the motor frame 30

Figure 8. The design of the main frame and fragrance storage design 31

Figure 9. Design of the Gyeongbokgung from the front 33

Figure 10. Design of the Gyeongbokgung from the side view 34

Figure 11. The design of the pump and control panel 36

Figure 12. Control block diagram of an automatic fragrance device 38

Figure 13. Automatic perfume device's program design. 40

Figure 14. The final looks of the automatic fragrance device. 42

Figure 15. The 3D drawing and final printing of Gyeongbokgung. 43

Figure 16. Automatic perfume device's coding. 45

Figure 17. Automatic perfume device's coding. 46

Figure 18. Automatic perfume device's coding. 47

Figure 19. Automatic perfume device's coding. 48

Figure 20. Automatic perfume device coding. 49

Figure 21. Automatic perfume device's coding. 50

Figure 22. Automatic perfume device's coding. 51

Figure 23. Automatic perfume device's coding. 52

Figure 24. The fragrance device pump control program. 53

Figure 25. Automatic perfume device's program design. 56

Figure 26. Program procedure for open data collection. 58

Figure 27. Experiment perfume arrangement in the box. 60

Figure 28. Fragrance Combination and manufacturing method. 61

Figure 29. The diagram of the fragrance evaluation program. 65

Figure 30. The diagram of recipe recommendation system. 68

Figure 31. Manual fragrance evaluation program for closed data. 78

Figure 32. Perceptron machine learning algorithm. 83

Figure 33. Perceptron machine learning algorithm. 84

Figure 34. Multi-perceptron machine learning algorithm. 85

Figure 35. Multi-perceptron machine learning algorithm. 86

Figure 36. Deep-multi-perceptron machine learning algorithm. 87

Figure 37. Deep-multi-perceptron machine learning algorithm. 88

Figure 38. Deep-multi-perceptron machine learning algorithm. 89

Figure 39. Perfume recipe recommendation system program. 90

Figure 40. Michael Edwards' Fragrance Wheel 2014. 91

초록보기

 본 논문은 어떤 특정한 향수를 개발하기 위해서는 특정한 향취의 향료를 조합하는 것을 오픈 데이터와 closed 데이터를 모으고, 이러한 자료를 인공지능 프로그램으로 추천하는 시스템을 구축하였다. 조향사들이 활약하는 시기에는 오랜 기간 도제식 훈련을 통해 어떤 향수를 만들려면 어떤 향수 원료가 들어가야 한다는 자료가 구축되어야 했지만, 인터넷이 활발하게 오픈되어 있는 현재는 인터넷으로부터 향수 원료에 대한 레시피를 구하여 빅데이터를 구축할 수 있게 되었다. 저자는 인터넷으로 약 54,000개의 오픈 데이터를 확보하여 오픈 데이터를 구축하였으며, 이러한 데이터를 바탕으로 시트러스 (Citrus) 향료 조합을 통해 1,026 클로스드(closed) 데이터를 구축하였다. 이러한 데이터를 마이크로소프트 엑셀에 정리를 하고, 파이썬을 이용하여 프로그램을 코딩하였다. 코딩을 통해 데이터를 퍼셉트론, 멀티 퍼셉트론, 딥 멀티 퍼셉트론 강화학습으로 시트러스 향료 조합에 필요한 레시피가 추천되도록 하였다. 본 논문을 통해 citrus-floral-sweet 향취 타입, citrus-fruity-fresh 타입, citrus-sweet-oriental 타입, sweet-citrus-musky 향취 타입, citrus-acidic-sweet 향취 타입, aldehyde-minty-sweet 향취 타입의 향수를 제조하는데 필요한 천연향료와 합성향료를 추천할 수 있었다.

추천서가 (다양한 추천 자료를 만나보세요)

권호기사보기

권호기사 목록 테이블로 기사명, 저자명, 페이지, 원문, 기사목차 순으로 되어있습니다.
기사명 저자명 페이지 원문 기사목차