입 추적을 이용한 로봇 원격 제어 시스템

Contents 1. INTRODUCTION 1 1.1 MOTIVATION 1 1.2 APPROACH AND CONTRIBUTIONS 3 1.3 OUTLINE 5 2. BACKGROUND AND RELATED RESEARCH 6 3. SYSTEM OVERVIEW 8 3.1 CONTROLLING END 9 3.2 ROBOT END 11 4. MOUTH TRACKING 13 4.1 ADABOOST ALGORITHM INTRODUCTION 13 4.2 GENTLE ADABOOST ALGORITHM 14 4.3 HAAR-LIKE FEATURES 15 4.3.1 Haar-like features introduction 15 4.3.2 Integral image 20 4.4 OVERVIEW OF ADABOOST MOUTH DETECTOR TRAINING 23 4.5 TRAINING OF DOUBLE HALF-MOUTH DETECTOR 26 4.5.1 What is double half-mouth detector? 26 4.5.1 Sample selection 28 4.5.2 Feature selection 30 4.5.3 Weak classifier 31 4.5.4 Train (stage) classifier by Gentle Adaboost algorithm 32 4.5.5 Cascade classifier 34 4.5.6 Multi-scale mouth detection 37 4.6 DETECTION POST PROCESSING 39 4.6.1 Analysis of multi-scale overlapping regions 39 4.6.2 Analysis of YCbCr color differentiation 41 4.6.3 Position speculate by Double half-mouth detectors 42 4.7 MOUTH TRACKING 43 4.7.1 The overall structure of mouth tracking system 43 4.7.2 K