Flash-type CMOS Inverter as Programmable Gaussian-like Synapse for Probabilistic Neural Networks
확률적 신경망을 위한 플래시형 CMOS 인버터 기반 프로그래머블 가우시안 유사 시냅스
- 주제(키워드) CMOS compatible anti-ambipolar transistor , Probabilistic Neural Network
- 발행기관 아주대학교 일반대학원
- 지도교수 Jang Hyun Kim
- 발행년도 2026
- 학위수여년월 2026. 8
- 학위명 석사
- 학과 및 전공 일반대학원 지능형반도체공학과
- 실제URI http://www.dcollection.net/handler/ajou/000000036543
- 본문언어 영어
- 저작권 아주대학교 논문은 저작권에 의해 보호받습니다.
목차
Section 1. Introduction 1
1.1 Concept and Advantages of Probabilistic Neural Networks 1
1.2 Necessity of CMOS Process-Based Gaussian-like synapse Development 7
1.3 Proposed Flash-Type CMOS Inverter Structure as a Gaussian-like synapse 12
Section 2. Device Fabrication and Measurement of Electrical Characteristics 15
2.1 Device Fabrication 15
2.2 Electrical Characterization 20
Section 3. Probabilistic Neural Network Simulation for ECG Signal Classification 29
3.1 ECG Signal Preprocessing 29
3.2 Gaussian Curve Modeling Based on Flash-Type CMOS Inverter Characteristics 32
3.3 Gaussian-like synapse-Based Representation of ECG FFT Features 35
3.4 Construction of the PNN Architecture Using Programmable Gaussian-like synapse 40
3.5 ECG Classification Results and Accuracy Analysis 43
Section 4. Conclusion 48
References 49
