| 일 | 월 | 화 | 수 | 목 | 금 | 토 |
|---|---|---|---|---|---|---|
| 1 | 2 | 3 | ||||
| 4 | 5 | 6 | 7 | 8 | 9 | 10 |
| 11 | 12 | 13 | 14 | 15 | 16 | 17 |
| 18 | 19 | 20 | 21 | 22 | 23 | 24 |
| 25 | 26 | 27 | 28 | 29 | 30 | 31 |
- Read
- Class
- atmega328
- stream
- memory
- SERIAL
- Pointer
- file access
- APP
- Gaussian
- aduino
- Filtering
- Contour
- flutter
- compare
- wpf
- UNO
- digitalRead
- sensor
- subpixel
- Android
- c++
- Encapusulation
- parameter
- edge
- Binary
- Unity
- mfc
- Gradient
- public
- Today
- Total
목록confidence (2)
폴크(FOLC)
1차 중앙 차분(Central Difference) - 3개의 샘플만을 이용해서 처리고차 차분 기반 수치 미분- 3개 / 5개 / 7개 / Forward / Backward 등 정규화 마스크 기반 미분 (Sobel/Prewitt/Scharr 등)- Sobel / Prewitt / Scharr / Roberts 함수 피팅 기반 (Local Regression / Polynomial)- Savitzky–Golay 미분 필터 / Local Quadratic Fitting / Spline Derivatives 미분 필터 함수 기반 (선형 시스템 해석 기반)- Gaussian Derivative (DoG) / Difference of Gaussians (DoG) / Derivative of Gaussian Ker..
Gradient 기반 신뢰도 스코어링(gradient-based edge confidence scoring)**은 검출된 edge가 진짜 edge인지 아닌지를 판단하거나 subpixel edge의 신뢰도를 정량화할 때 매우 유용한 기법으로, CD 측정, 자동 edge 선택, 노이즈 필터링 등에 적용된다.Gradient magnitude : edge의 "강도(strength)" 또는 "확신도(confidence)"로 해석 가능 강한 edge : 큰 gradient → 확실한 밝기 변화 → 진짜 edge일 가능성 높음 약한 edge : 작은 gradient → noise, blur, 혹은 경계 아님 Confidence = |I(x+1) - I(x-1)| / 2 = |Gradient| 스코어링 기준값 예시≥..