SlideShare uma empresa Scribd logo
1 de 37
Baixar para ler offline
응용개발사업부
HDR & Tone Mapping (1/36)
2021/03/19
HDR & Tone Mapping
HDR 기술의 소개부터 딥러닝 기반 최신 톤 매핑 기법까지
Tech Seminar
Presented by 응용개발사업부 이명규
응용개발사업부
HDR & Tone Mapping (2/36)
I N D E X
01
02
03
Introduction
HDR & Tone
Mapping
Featured Papers
응용개발사업부
HDR & Tone Mapping (3/36)
Introduction
Part 01
1. HDR Image
2. HDR In Various Fields
응용개발사업부
HDR & Tone Mapping (4/36)
HDR Image
1-1
• HDR: “Content with a Wider Range of Brightness and Color”
• HDR을 잘 표현하기 위해서는 새로운 디스플레이 장치가 필요
• 기존 콘텐츠: 최대 100nit 밝기, 709 Gamut*
• HDR 콘텐츠: 최대 10,000nit 밝기, 2020 Gamut
HDR LDR
*Gammut: 색상과 채도 (밝기와 대비를 의미하는 Gamma와 구분)
High-Dynamic Range (HDR) Demystified
Object
Approx.
Luminance
(nits)
Sun 1,600,000,000
Arc Lamp 150,000,000
Maximum
Visual Tolerance
50,000
Cloud (Sunny Day) 35,000
2016 UHD TV 800~1,000
Typical Computer Screen 100~300
White Paper
Under the Lamp
50
Night Sky 0.001
Threshold of Vision 0.000003
응용개발사업부
HDR & Tone Mapping (5/36)
HDR In Various Fields
1-2
VOD Service Gaming Photography
High Dynamic Range Imaging 기술 및 최근 동향 (The Korean Information Display Society, 2019)
• HDR Imaging?
• 입력 영상에 대해 노출 정도를 다르게 설정 후 취득한 여러 이미지를 합성함으로써 Dynamic Range 증강
• 카메라에 의해 취득된 HDR 영상을 LCD, OLED 등의 디스플레이에서 정확히 표현할 수 있도록 하는 것
응용개발사업부
HDR & Tone Mapping (6/36)
HDR In Various Fields
1-2
High Dynamic Range Imaging Technology
응용개발사업부
HDR & Tone Mapping (7/36)
HDR & Tone
Mapping
Part 02
1. What is HDR?
2. What is Tone Mapping?
응용개발사업부
HDR & Tone Mapping (8/36)
↳
What is HDR?
2-1
Dynamic Range of Human Eye
Use Cyclone® V SoC FPGA to Create Real-time HDR Video(Intel), Investigation on the Use of HDR Images for Cultural Heritage Documentation
• Contrast Ratio란?
• 디스플레이가 출력 가능한 가장 밝은 색(white)과
어두운 색(Black)의 휘도(Luminance) 명암 비율
응용개발사업부
HDR & Tone Mapping (9/36)
↳
What is HDR?
2-1
Dynamic Range of Human Eye
'Light Adaptation'(Vision Models for High Dynamic Range and Wide Colour Gamut Imaging, 2020), “High Quality High Dynamic Range Imaging”
Cone Function (추상체 수용기 감지범위)
No
Moon
Moonlight
(Fullmoon)
Early
Twilight
Store,
Office
Outdoors
(Sunny)
Starlight
𝟏𝟎−𝟔
𝟏𝟎−𝟓 𝟏𝟎−𝟒
𝟏𝟎−𝟑
𝟏𝟎−𝟐
𝟏𝟎−𝟏 𝟏 𝟏𝟎𝟏
𝟏𝟎𝟐
𝟏𝟎𝟑
𝟏𝟎𝟒
𝟏𝟎𝟓 𝟏𝟎𝟔
Luminance
(𝒄𝒅/𝒎𝟐
)
Rod Function (간상체 수용기 감지범위)
HDR Display
Normal Display
Image Capturing
(Loss Dynamic Range)
Domain of Human Vision:
≈ 𝟏𝟎−𝟒
~ ≈ 𝟏𝟎𝟔
Conventional Image:
−𝟏 ~ ≈ 𝟏𝟎𝟐
“LDR Image”
(8-bit Integer [0~255])
“HDR Image”
(32-bit Floating Point)
HDR Imaging
(Recover lost Dynamic Range)
응용개발사업부
HDR & Tone Mapping (10/36)
↳
What is HDR?
2-1
LDR VS. HDR
High Dynamic Range Imaging 기술 및 최근 동향
• Contrast Ratio에 따른 신호 구분
• 명암비가 1000:1보다 작은 경우: LDR(Low Dynamic Range) 또는 SDR(Standard Dynamic Range)
• 명암비가 1000:1~100,000:1인 경우: EDR (Enhanced Dynamic Range)
• 명암비가 100,000:1보다 큰 경우: HDR(High Dynamic Range)
• HDR 영상의 활용을 위해서는 디스플레이의 영상 신호처리 기술과 함께
색 재현율에 대한 기본적인 이해 필요
Today’s Topic
응용개발사업부
HDR & Tone Mapping (11/36)
↳
What is HDR?
2-1
LDR VS. HDR
[0107 박민근] 쉽게 배우는 hdr과 톤맵핑
• LDR(SDR)
➢0~255로 256단계 색상 표현 가능 (𝟐𝟓𝟔𝟑
= 𝟏𝟔, 𝟕𝟕𝟕, 𝟐𝟏𝟔개 색 표현 가능)
➢픽셀당 4Byte씩 총 32-bit 사용 (R: 8-bit, G: 8-bit, B: 8-bit, A: 8-bit)
• HDR(OpenEXR format)
➢색상 당 16-bit Floating Point로 표현
➢채널 당 부호 1-bit, 가수에 10-bit, 지수에 5-bit 사용 (알파채널 포함 총 64-bit 버퍼 사용)
➢약 𝟒. 𝟒 × 𝟏𝟎𝟏𝟐
= 4.4e+12 색상 표현 가능 (≈ 𝟒, 𝟒𝟎𝟎, 𝟎𝟎𝟎, 𝟎𝟎𝟎, 𝟎𝟎𝟎개)
응용개발사업부
HDR & Tone Mapping (12/36)
↳
What is HDR?
2-1
LDR VS. HDR
High-Dynamic Range (HDR) Demystified, [디스플레이 톺아보기] ㉚ HDR(High Dynamic Range)의 이해
*SDR: Standard Dynamic Range
(30 nit laptops in low power mode, 600 nit HDTVs in vivid mode)
Real World Mastered for SDR* Mastered for HDR
응용개발사업부
HDR & Tone Mapping (13/36)
↳
What is HDR?
2-1
Color Bit
[디스플레이 톺아보기] ㉙ 디스플레이 색심도(Color Depth)의 이해, Converting Color Depth | Color
8-bit
3-bit 8-bit 24-bit
10-bit
응용개발사업부
HDR & Tone Mapping (14/36)
↳
What is HDR?
2-1
HDR Standard Formats
[디스플레이 톺아보기] ㉚ HDR(High Dynamic Range)의 이해
HDR10 Dolby Vision
Color Gamut
(색역)
Color Depth
(색심도)
10bit 12bit
Peak Luminance
(최대밝기)
1000nits 10000nits
Meta Data
Static
(콘텐츠별 설정)
Dynamic
(프레임별 설정)
BT.2020
응용개발사업부
HDR & Tone Mapping (15/36)
↳
What is HDR?
2-1
HDR Standard Formats
High-dynamic-range video
HDR10 HDR10+ Dolby Vision HLG10
CTA Samsung Dolby NHK and BBC
2015 2017 2014 2015
Free
Free (for content company),
Yearly license (for manufacturer)
Proprietary Free
Static
(SMPTE ST 2086, MaxFALL, MaxCLL)
Dynamic
Dynamic
(Dolby Vision L0, L1, L2 trim, L8 trim)
None
PQ PQ PQ (Not always) HLG
10 bit 10 bit (or more) 10 bit or 12 bit 10 bit
Technical
limit
10,000 nits 10,000 nits 10,000 nits Variable
Contents
No rules
1,000 - 4,000 nits (common)
No rules
1,000 - 4,000 nits (common)
(At least 1,000 nits)
4,000 nits common
1,000 nits common
Technical
limit
Rec. 2020 Rec. 2020 Rec. 2020 Rec. 2020
Contents DCI-P3 (common) DCI-P3 (common) At least DCI-P3 DCI-P3 (common)
None HDR10
It depends on the profile used:
- No compatibility
- SDR
- HDR10
- HLG
- Ultra HD Blu-Ray
- UHD-TV (Rec.2020)
- SDR (Rec.709) with color
distortion
Transfer function
Bit Depth
Peak 
luminance
Color
primaries
Backward
compatibility
Metadata
Developed by
Year
Cost
Technical
charasteristics
응용개발사업부
HDR & Tone Mapping (16/36)
↳
What is HDR?
2-1
모니터의 영상 신호 처리
High Dynamic Range Imaging 기술 및 최근 동향
• 디스플레이는 인간의 시각적 특성을 고려해 신호 처리
• 베버의 법칙에 따라, 정해진 최대 표현량을 정확히 고려해 최적의 화질을 재생할 수 있는
비선형 관계 이해가 중요
• 단순히 선형적으로 빛의 밝기를 표현하면 Posterization 현상이 발생하며,
이를 해결하기 위해 감마 보정 기법 사용
• HDR Imaging은 기존 디스플레이보다 높은 10~12-bit 범위에서 영상을 표현하므로,
주어진 bit depth 내에서 입력/출력 신호 간 관계를 적절하게 정의해야 함
• HDR Imaging은 또한 기존에 사용되던 Standard RGB 대비 170% 증가된
색역(e.g. BT Rec.2020)을 지원해야 함.
응용개발사업부
HDR & Tone Mapping (17/36)
↳
• “인간의 감각은 선형적이지 않다”
➢초기에 받은 자극의 정도에 따라 나중에 받는 자극의 수용폭이 달라진다는 생리학 이론
Weber’s Law
손바닥에 100g의 무게부터 조금씩 무게를 늘려 나갔을 때 102g에서 최초로 무게가 다르다는 것을 느낄 수가 있었고,
200g의 물건을 올려놓았을 때는 204g에서 최초로 무게가 다르다는 것을 느낄 수 있었다.
이 실험에서 베버는 최초로 차이를 느낄 수 있을 때의 자극의 증가량 2g, 4g(R 절대판별역)과 처음 올려놓은 표준자극
100g과 200g(R)의 비(R/R상대판별역)는 항상 비례적으로 일정하다는 것을 발견했다.
이렇게 감각으로 구별할 수 있는 한계는 물리적인양의 차이가 아니고, 그 비율관계에 의해 결정된다는 것이다.
베버-페흐너의 법칙(Weber-Fechner’s law), UX디자이너가 알아야 할 심리학 법칙 5가지
What is HDR?
2-1
응용개발사업부
HDR & Tone Mapping (18/36)
↳
Gamma Correction
머신 비전 ISP – 17. Gammar Corection, 감마보정 Gamma Correction
• “모니터는 실제보다 어둡게 빛을 표현한다!”
• 사람의 눈은 어두운 환경에서 작은 밝기의 변화에도 민감하게 반응
(반면 밝은 환경에서는 작은 밝기의 변화에 둔감하게 반응)
• “인간의 시각은 비선형적이므로, 굳이 인간이 잘 느끼지 못하는 부분까지 정밀하게
계산할 필요는 없다”
𝒇 𝒙 = 𝑮𝒂𝒊𝒏 × 𝑿𝒈𝒂𝒎𝒎𝒂
+ 𝒐𝒇𝒇𝒔𝒆𝒕
Nonlinear
Transfer
Function
Intensity
(HDR)
Transferred
Intensity
(Display)
일반적인
CRT 모니터의 감마
What is HDR?
2-1
응용개발사업부
HDR & Tone Mapping (19/36)
↳
Gamma Correction
Lighting Shading by John Hable, 이미지 파일 감마 (Image File Gamma)와 디스플레이 감마 (Display Gamma)
• 이미지는 일반적으로 1/2.2로 감마를 적용한 상태에서 저장
• 디스플레이 출력 시 올바른 색상 표현을 위해 저장 시의 Gamma를 상쇄해야 함
• 올바른 HDR 결과를 얻기 위해서는 감마 보정 작업을 거쳐야 한다!
What is HDR?
2-1
응용개발사업부
HDR & Tone Mapping (20/36)
↳
What is Tone Mapping?
2-2
• 따라서 HDR 포맷을 디스플레이 출력이 가능한 휘도 범위로 변환 처리해야 함.
• 톤 매핑은 HDR 범위를 인간 색 인지에 기반해 Display에서 지원하는 범위로 맞추는 작업
• Tone Mapping 기법의 종류
• HDR to HDR
• HDR to LDR
• LDR to HDR (“Inverse Tone Mapping”)
Tone Mapping
• HDR(FP16)
➢ 색상 당 16-bit Floating Point로 표현
➢ 채널 당 부호 1-bit, 가수에 10-bit, 지수에 5-bit 사용 (알파채널 포함 총 64-bit 버퍼 사용)
➢ 약 𝟒. 𝟒 × 𝟏𝟎𝟏𝟐
= 4.4e+12 색상 표현 가능 (≈ 𝟒, 𝟒𝟎𝟎, 𝟎𝟎𝟎, 𝟎𝟎𝟎, 𝟎𝟎𝟎)
“아무리 실수 범위에서 넓게 계산해도
일반 모니터는 4Byte RGB만 출력 가능”
Today’s Topic
응용개발사업부
HDR & Tone Mapping (21/36)
↳
What is Tone Mapping?
2-2
Tone Mapping
SDR
Content
SDR
Display
SDR
Experience
SDR
Content
HDR
Display
SDR
Experience
HDR
Content
HDR
Display
HDR
Experience
HDR
Content
SDR
Display
Bad
Experience
HDR
Content
HDR▶SDR
Tone Mapping
SDR
Display
SDR
Experience
High-Dynamic Range (HDR) Demystified
응용개발사업부
HDR & Tone Mapping (22/36)
↳
What is Tone Mapping?
2-2
• HDR Imaging에는 비싼 장비와 상당한 연산 시간이 필요
➢ 따라서 LDR 이미지만으로 Real World Luminance를 추정하는 역 톤 매핑 기법의 필요성 대두
• 톤 매핑 과정에서 Linear RGB를 CIE 1931 XYZ Color Space로 변환
➢ 𝑿 = 𝟎. 𝟒𝟏𝟐𝑹𝒘 + 𝟎. 𝟑𝟓𝟕𝑮𝒘 + 𝟎. 𝟏𝟖𝑩𝒘
𝒀 = 𝑳𝑾 = 𝟎. 𝟐𝟏𝟑𝑹𝒘 + 𝟎. 𝟕𝟏𝟓𝑮𝒘 + 𝟎. 𝟎𝟕𝟐𝑩𝒘, where 𝐿𝑊 is real world luminance(≈HDR)
𝒁 = 𝟎. 𝟎𝟏𝟗𝑹𝒘 + 𝟎. 𝟏𝟏𝟗𝑮𝒘 + 𝟎. 𝟗𝟓𝑩𝒘
➢ XYZ Color Space는 인간의 색채 인지 연구를 바탕으로 만들어진 색 공간 (따라서 다른 색공간의 기본이 됨)
• 톤 매핑 과정에서 압축된 정보의 손실이 발생하므로 역 톤 매핑은 ill-posed problem에 해당
• 역 톤 매핑은 𝑹𝒅, 𝑮𝒅, 𝑩𝒅만 주어지므로 Real World Color(𝑹𝒘, 𝑮𝒘, 𝑩𝒘)로 돌아가는 변환인 𝑳𝑾을 알 수 없음
Inverse Tone Mapping
"Inverse tone mapping“, ITU-R Recommendation BT.709 RGB, CIE 1931 color space, high Dynamic Range Imaging: Acquisition, Display and Image-Based Lighting
𝑹𝒅
𝑮𝒅
𝑩𝒅
=
𝑳𝒅
𝑹𝒘
𝑳𝒘
𝑳𝒅
𝑮𝒘
𝑳𝒘
𝑳𝒅
𝑩𝒘
𝑳𝒘
▲ Tone Mapping
𝑹𝒘
𝑮𝒘
𝑩𝒘
=
𝑳𝒘
𝑹𝒅
𝑳𝒅
𝑳𝒘
𝑮𝒅
𝑳𝒅
𝑳𝒘
𝑩𝒅
𝑳𝒅
▲ Inverse Tone Mapping
Compressed
Colors
World
Colors
???
응용개발사업부
HDR & Tone Mapping (23/36)
↳
What is Tone Mapping?
2-2
• Previous Approaches
➢“The Reproduction of Specular Highlights on High Dynamic Range Displays”
✓ 출력 디스플레이 특성을 고려해 Dynamic Range를 적응적으로 조절하는 방법
✓ 빛이 방출되거나 반사되는 영역에 대해 더 많은 정보를 부여해 표현력 향상
➢“Ldr2Hdr: on-the-fly reverse tone mapping of legacy video and photographs”
✓ 입력 영상 내에서 상위 계조를 갖는 영역 검출 후 더 많은 정보를 부여해 표현력 향상
✓ 빛 분포 고려 후 픽셀 표현 가능 범위를 넘어 saturation된 상위 계조에 집중
✓“Physiological inverse tone mapping based on retina response”
✓ 인간의 시각적 특성을 반영한 Perceptual Brightness를 정의해 입출력 영상 신호 간
상관 관계를 적응적으로 정의
Inverse Tone Mapping: Method
High Dynamic Range Imaging 기술 및 최근 동향 (The Korean Information Display Society, 2019)
응용개발사업부
HDR & Tone Mapping (24/36)
↳
What is Tone Mapping?
2-2
• Deep-Learning Based Approaches
➢“HDR image reconstruction from a single exposure using deep CNNs”
✓ 입력 영상에 대해 상위/하위 계조를 나눈 후 하위 계조는 원본 신호를,
상위 계조는 CNN을 통해 추론된 신호를 사용(단, 전체적으로 어두운 영상에서만 잘 작동)
➢“Deep Recursive HDRI: Inverse Tone Mapping using Generative Adversarial Networks”
✓ GAN을 활용해 성능을 향상(L1-Loss, Adversarial Loss, U-Net, PatchGAN)
Inverse Tone Mapping: Method
High Dynamic Range Imaging 기술 및 최근 동향 (The Korean Information Display Society, 2019)
응용개발사업부
HDR & Tone Mapping (25/36)
Featured Papers
Part 03
1. HDRCNN
(SIGGRAPH ASIA 2017)
2. SingleHDR
(CVPR 2020)
3. LDR2HDR
(SIGGRAPH 2020)
응용개발사업부
HDR & Tone Mapping (26/36)
↳
HDRCNN
3-1
"HDR image reconstruction from a single exposure using deep CNNs"
(SIGGRAPH ASIA 2017, GABRIEL EILERTSEN et al.) [Paper, Code]
https://arxiv.org/abs/1710.07480
“Highlight Information을 유지하면서(초록 박스)
Saturate된 색상 영역의 Color Detail 복원”
응용개발사업부
HDR & Tone Mapping (27/36)
↳
Convert to
logarithmic HDR
Domain
Performs Bilinear
Up-sampling
Performs
Addition
Predicted HDR image
in log domain
HDRCNN
3-1
Network Structure
https://arxiv.org/abs/1710.07480
응용개발사업부
HDR & Tone Mapping (28/36)
↳
HDRCNN
3-1
Problem Formulation & Loss Function
https://arxiv.org/abs/1710.07480, Modeling the Space of Camera Response Functions
𝓛 ෝ
𝒚, 𝑯 =
𝟏
𝟑𝑵
෍
𝒊,𝒄
𝜶𝒊(ෝ
𝒚𝒊,𝒄 − 𝐥𝐨𝐠(𝑯𝒊,𝒄 + 𝝐))
𝟐
𝑯𝒊,𝒄 ∈ ℝ+
# of Pixels
Linear
GT HDR
Predicted
Log HDR
▲ 201 Real-world Response Functions Database(“DoRF”)
෡
𝑯𝒊,𝒄 = 𝟏 − 𝜶𝒊 𝒇−𝟏 𝑫𝒊,𝒄 + 𝜶𝒊𝐞𝐱𝐩(ෝ
𝒚𝒊, 𝒄)
Final Reconstructed
HDR pixels
(where 𝒊=spatial index,
𝒄=channel)
Input
LDR Pixel
Predicted output
(log domain)
Inverse Camera Function
(transform input to linear domain)
Blending
Factor
𝜶𝒊 =
𝐦𝐚𝐱(𝟎,𝒎𝒂𝒙𝒄 𝑫𝒊,𝒄 −𝝉)
𝟏−𝝉
, where 𝝉 =0.95
(하이라이트 부근의 Banding artifact를 방지하기 위한 factor)
응용개발사업부
HDR & Tone Mapping (29/36)
↳
SingleHDR
3-2
"Single-Image HDR Reconstruction by Learning to Reverse the
Camera Pipeline" (CVPR 2020, Yu-Lun Liu et al.) [Paper, Code]
https://arxiv.org/abs/2004.01179
“카메라 파이프라인을 역으로 모델링해 학습함으로써
소실된 디테일 정보 복원”
응용개발사업부
HDR & Tone Mapping (30/36)
↳
SingleHDR
3-2
Network Structure
https://arxiv.org/abs/2004.01179
▲ LDR Image formation pipeline
▲ Proposed Method
응용개발사업부
HDR & Tone Mapping (31/36)
↳
SingleHDR
3-2
Network Structure
https://arxiv.org/abs/2004.01179
▲ Linearization Network
▲ Hallucination Network
응용개발사업부
HDR & Tone Mapping (32/36)
↳
LDR2HDR
3-3
"Single Image HDR Reconstruction Using a CNN with Masked Features
and Perceptual Loss" (SIGGRAPH 2020, MARCEL SANTANA SANTOS et al.) [Paper, Code]
https://people.engr.tamu.edu/nimak/Data/SIGGRAPH20_HDR.pdf
“높은 노출로 인해 밝게 타 버린 텍스쳐 영역도 잘 복원”
응용개발사업부
HDR & Tone Mapping (33/36)
↳
LDR2HDR
3-3
Demo Images
https://people.engr.tamu.edu/nimak/Data/SIGGRAPH20_HDR.pdf
응용개발사업부
HDR & Tone Mapping (34/36)
↳
LDR2HDR
3-3
Network Structure
https://people.engr.tamu.edu/nimak/Data/SIGGRAPH20_HDR.pdf
일반적인 문제와는 달리 타 버린(saturated areas) LDR 영역에
대해서도 잘 작동시키기 위해 Feature Masking 기법을 제안
응용개발사업부
HDR & Tone Mapping (35/36)
↳
LDR2HDR
3-3
Network Structure: Feature Masking
https://people.engr.tamu.edu/nimak/Data/SIGGRAPH20_HDR.pdf
𝒁𝒍 = 𝑿𝒍⨀𝑴𝒍
𝑿𝒍+𝟏 = 𝝓𝒍(𝑾𝒍 ∗ 𝒁𝒍 + 𝒃𝒍)
𝑿𝒍 ∈ ℝ𝑯×𝑾×𝑪
, 𝑴𝒍 ∈ 𝟎, 𝟏 𝑯×𝑾×𝑪
𝑊𝑙 and 𝑏𝑙 is weight and bias of the current layer
응용개발사업부
HDR & Tone Mapping (36/36)
↳
LDR2HDR
3-3
Loss Function
https://people.engr.tamu.edu/nimak/Data/SIGGRAPH20_HDR.pdf
𝑳 = 𝝀𝟏𝑳𝒓 + 𝝀𝟐𝑳𝒑, 𝒘𝒉𝒆𝒓𝒆 𝝀𝟏 = 𝟔. 𝟎 𝒂𝒏𝒅 𝝀𝟐 = 𝟏. 𝟎
Reconstruction
Loss
Perception
Loss
𝑳𝒓 = (𝟏 − 𝑴 ⊙ (෡
𝒀 − 𝐥𝐨𝐠(𝑯 + 𝟏)) 𝟏
Loss가 Saturate된 영역에서
계산되도록 마스킹
𝑳𝒑 = 𝝀𝟑𝑳𝝊 + 𝝀𝟒𝑳𝒔
𝑳𝝊 = ෍
𝒍
𝝓𝒍 𝓣 ෩
𝑯 − 𝝓𝒍(𝓣(𝑯))
𝟏
𝑳𝑺 = ෍
𝒍
𝑮𝒍 𝓣 ෩
𝑯 − 𝑮𝒍(𝓣(𝑯))
𝟏
𝑮𝒍 𝑿 =
𝟏
𝑲𝒍
𝝓𝒍 𝑿 𝑻𝝓𝒍 𝑿 .
(𝑮𝒍 is Gram matrix of the feature layer 𝒍.)
응용개발사업부
HDR & Tone Mapping (37/36)
Thank you for Watching.
brstar96@espresomedia.com (이명규)

Mais conteúdo relacionado

Mais procurados

Color models in Digitel image processing
Color models in Digitel image processingColor models in Digitel image processing
Color models in Digitel image processingAryan Shivhare
 
Introduction to HEVC
Introduction to HEVCIntroduction to HEVC
Introduction to HEVCYoss Cohen
 
Digital image processing ppt
Digital image processing pptDigital image processing ppt
Digital image processing pptkhanam22
 
Advance image processing
Advance image processingAdvance image processing
Advance image processingAAKANKSHA JAIN
 
Compression: Images (JPEG)
Compression: Images (JPEG)Compression: Images (JPEG)
Compression: Images (JPEG)danishrafiq
 
Digital image processing
Digital image processingDigital image processing
Digital image processingABIRAMI M
 
VVC tutorial at ICME 2020 together with Benjamin Bross
VVC tutorial at ICME 2020 together with Benjamin BrossVVC tutorial at ICME 2020 together with Benjamin Bross
VVC tutorial at ICME 2020 together with Benjamin BrossMathias Wien
 
Image compression
Image compressionImage compression
Image compressionPREEYANKAV
 
An Introduction to Video Principles-Part 2
An Introduction to Video Principles-Part 2An Introduction to Video Principles-Part 2
An Introduction to Video Principles-Part 2Dr. Mohieddin Moradi
 
Fidelity criteria in image compression
Fidelity criteria in image compressionFidelity criteria in image compression
Fidelity criteria in image compressionKadamPawan
 
Iain Richardson: An Introduction to Video Compression
Iain Richardson: An Introduction to Video CompressionIain Richardson: An Introduction to Video Compression
Iain Richardson: An Introduction to Video CompressionIain Richardson
 
Image compression in digital image processing
Image compression in digital image processingImage compression in digital image processing
Image compression in digital image processingDHIVYADEVAKI
 
Introduction to Video Compression Techniques - Anurag Jain
Introduction to Video Compression Techniques - Anurag JainIntroduction to Video Compression Techniques - Anurag Jain
Introduction to Video Compression Techniques - Anurag JainVideoguy
 
Color image processing
Color image processingColor image processing
Color image processingrmsurya
 
Compression: Video Compression (MPEG and others)
Compression: Video Compression (MPEG and others)Compression: Video Compression (MPEG and others)
Compression: Video Compression (MPEG and others)danishrafiq
 

Mais procurados (20)

Color models in Digitel image processing
Color models in Digitel image processingColor models in Digitel image processing
Color models in Digitel image processing
 
Introduction to HEVC
Introduction to HEVCIntroduction to HEVC
Introduction to HEVC
 
Digital image processing ppt
Digital image processing pptDigital image processing ppt
Digital image processing ppt
 
Advance image processing
Advance image processingAdvance image processing
Advance image processing
 
Compression: Images (JPEG)
Compression: Images (JPEG)Compression: Images (JPEG)
Compression: Images (JPEG)
 
Digital image processing
Digital image processingDigital image processing
Digital image processing
 
VVC tutorial at ICME 2020 together with Benjamin Bross
VVC tutorial at ICME 2020 together with Benjamin BrossVVC tutorial at ICME 2020 together with Benjamin Bross
VVC tutorial at ICME 2020 together with Benjamin Bross
 
HDR and WCG Principles-Part 6
HDR and WCG Principles-Part 6HDR and WCG Principles-Part 6
HDR and WCG Principles-Part 6
 
Image compression
Image compressionImage compression
Image compression
 
An Introduction to Video Principles-Part 2
An Introduction to Video Principles-Part 2An Introduction to Video Principles-Part 2
An Introduction to Video Principles-Part 2
 
Fidelity criteria in image compression
Fidelity criteria in image compressionFidelity criteria in image compression
Fidelity criteria in image compression
 
Iain Richardson: An Introduction to Video Compression
Iain Richardson: An Introduction to Video CompressionIain Richardson: An Introduction to Video Compression
Iain Richardson: An Introduction to Video Compression
 
Image compression in digital image processing
Image compression in digital image processingImage compression in digital image processing
Image compression in digital image processing
 
Broadcast Lens Technology Part 1
Broadcast Lens Technology Part 1Broadcast Lens Technology Part 1
Broadcast Lens Technology Part 1
 
HDR and WCG Principles-Part 3
HDR and WCG Principles-Part 3HDR and WCG Principles-Part 3
HDR and WCG Principles-Part 3
 
Image Compression
Image CompressionImage Compression
Image Compression
 
85 videocompress
85 videocompress85 videocompress
85 videocompress
 
Introduction to Video Compression Techniques - Anurag Jain
Introduction to Video Compression Techniques - Anurag JainIntroduction to Video Compression Techniques - Anurag Jain
Introduction to Video Compression Techniques - Anurag Jain
 
Color image processing
Color image processingColor image processing
Color image processing
 
Compression: Video Compression (MPEG and others)
Compression: Video Compression (MPEG and others)Compression: Video Compression (MPEG and others)
Compression: Video Compression (MPEG and others)
 

Semelhante a HDR and Tone Mapping Seminar Guide

HDR Insights Article 2 : PQ and HLG transfer functions for HDR
HDR Insights Article 2 : PQ and HLG transfer functions for HDRHDR Insights Article 2 : PQ and HLG transfer functions for HDR
HDR Insights Article 2 : PQ and HLG transfer functions for HDRVeneraTech
 
HDR Insights Article 3: Understanding HDR Tone Mapping
HDR Insights Article 3: Understanding HDR Tone MappingHDR Insights Article 3: Understanding HDR Tone Mapping
HDR Insights Article 3: Understanding HDR Tone MappingVeneraTech
 
Cis660 primer hdr_eric_cheng
Cis660 primer hdr_eric_chengCis660 primer hdr_eric_cheng
Cis660 primer hdr_eric_chengEric Cheng
 
HDR and WCG Video Broadcasting Considerations.pdf
HDR and WCG Video Broadcasting Considerations.pdfHDR and WCG Video Broadcasting Considerations.pdf
HDR and WCG Video Broadcasting Considerations.pdfssuserc5a4dd
 
Hdr Meets Black And White 2
Hdr Meets Black And White 2 Hdr Meets Black And White 2
Hdr Meets Black And White 2 Francesco Carucci
 
Latest Technologies in Production & Broadcasting
Latest  Technologies in Production & BroadcastingLatest  Technologies in Production & Broadcasting
Latest Technologies in Production & BroadcastingDr. Mohieddin Moradi
 
HDR Displays Note
HDR Displays NoteHDR Displays Note
HDR Displays NoteJoe Miseli
 
High Dynamic Range Imaging- A Review
High Dynamic Range Imaging- A ReviewHigh Dynamic Range Imaging- A Review
High Dynamic Range Imaging- A ReviewCSCJournals
 
“CMOS Image Sensors: A Guide to Building the Eyes of a Vision System,” a Pres...
“CMOS Image Sensors: A Guide to Building the Eyes of a Vision System,” a Pres...“CMOS Image Sensors: A Guide to Building the Eyes of a Vision System,” a Pres...
“CMOS Image Sensors: A Guide to Building the Eyes of a Vision System,” a Pres...Edge AI and Vision Alliance
 
Shooting High Dynamic Range
Shooting High Dynamic RangeShooting High Dynamic Range
Shooting High Dynamic RangeJessica Young
 
HDR Photography
HDR PhotographyHDR Photography
HDR PhotographyAmit Dash
 
20200509 sid china digital optics and digital modulation_v5.0
20200509 sid china digital optics and digital modulation_v5.020200509 sid china digital optics and digital modulation_v5.0
20200509 sid china digital optics and digital modulation_v5.0Chun-Wei Tsai
 
Recent Progress on Single-Image Super-Resolution
Recent Progress on Single-Image Super-ResolutionRecent Progress on Single-Image Super-Resolution
Recent Progress on Single-Image Super-ResolutionHiroto Honda
 
Pulse Estimation
Pulse EstimationPulse Estimation
Pulse EstimationSahil Shah
 
Iaetsd a novel approach to assess the quality of tone mapped
Iaetsd a novel approach to assess the quality of tone mappedIaetsd a novel approach to assess the quality of tone mapped
Iaetsd a novel approach to assess the quality of tone mappedIaetsd Iaetsd
 
Freedom in Lighting Design by Tuning the CCT with LEDs, LpS 2015, Bregenz
Freedom in Lighting Design by Tuning the CCT with LEDs, LpS 2015, BregenzFreedom in Lighting Design by Tuning the CCT with LEDs, LpS 2015, Bregenz
Freedom in Lighting Design by Tuning the CCT with LEDs, LpS 2015, BregenzWojtek Cieplik
 

Semelhante a HDR and Tone Mapping Seminar Guide (20)

HDR Insights Article 2 : PQ and HLG transfer functions for HDR
HDR Insights Article 2 : PQ and HLG transfer functions for HDRHDR Insights Article 2 : PQ and HLG transfer functions for HDR
HDR Insights Article 2 : PQ and HLG transfer functions for HDR
 
HDR Insights Article 3: Understanding HDR Tone Mapping
HDR Insights Article 3: Understanding HDR Tone MappingHDR Insights Article 3: Understanding HDR Tone Mapping
HDR Insights Article 3: Understanding HDR Tone Mapping
 
Cis660 primer hdr_eric_cheng
Cis660 primer hdr_eric_chengCis660 primer hdr_eric_cheng
Cis660 primer hdr_eric_cheng
 
HDR and WCG Video Broadcasting Considerations.pdf
HDR and WCG Video Broadcasting Considerations.pdfHDR and WCG Video Broadcasting Considerations.pdf
HDR and WCG Video Broadcasting Considerations.pdf
 
Hdr Meets Black And White 2
Hdr Meets Black And White 2 Hdr Meets Black And White 2
Hdr Meets Black And White 2
 
Latest Technologies in Production & Broadcasting
Latest  Technologies in Production & BroadcastingLatest  Technologies in Production & Broadcasting
Latest Technologies in Production & Broadcasting
 
HDR Displays Note
HDR Displays NoteHDR Displays Note
HDR Displays Note
 
High Dynamic Range Imaging- A Review
High Dynamic Range Imaging- A ReviewHigh Dynamic Range Imaging- A Review
High Dynamic Range Imaging- A Review
 
“CMOS Image Sensors: A Guide to Building the Eyes of a Vision System,” a Pres...
“CMOS Image Sensors: A Guide to Building the Eyes of a Vision System,” a Pres...“CMOS Image Sensors: A Guide to Building the Eyes of a Vision System,” a Pres...
“CMOS Image Sensors: A Guide to Building the Eyes of a Vision System,” a Pres...
 
Shooting High Dynamic Range
Shooting High Dynamic RangeShooting High Dynamic Range
Shooting High Dynamic Range
 
HDR Photography
HDR PhotographyHDR Photography
HDR Photography
 
HDR Workshop
HDR WorkshopHDR Workshop
HDR Workshop
 
20200509 sid china digital optics and digital modulation_v5.0
20200509 sid china digital optics and digital modulation_v5.020200509 sid china digital optics and digital modulation_v5.0
20200509 sid china digital optics and digital modulation_v5.0
 
Recent Progress on Single-Image Super-Resolution
Recent Progress on Single-Image Super-ResolutionRecent Progress on Single-Image Super-Resolution
Recent Progress on Single-Image Super-Resolution
 
Pulse Estimation
Pulse EstimationPulse Estimation
Pulse Estimation
 
Iaetsd a novel approach to assess the quality of tone mapped
Iaetsd a novel approach to assess the quality of tone mappedIaetsd a novel approach to assess the quality of tone mapped
Iaetsd a novel approach to assess the quality of tone mapped
 
Freedom in Lighting Design by Tuning the CCT with LEDs, LpS 2015, Bregenz
Freedom in Lighting Design by Tuning the CCT with LEDs, LpS 2015, BregenzFreedom in Lighting Design by Tuning the CCT with LEDs, LpS 2015, Bregenz
Freedom in Lighting Design by Tuning the CCT with LEDs, LpS 2015, Bregenz
 
Digital File Formats
Digital File Formats Digital File Formats
Digital File Formats
 
4K Display Technology
4K Display Technology4K Display Technology
4K Display Technology
 
Scz 2370
Scz 2370Scz 2370
Scz 2370
 

Mais de MYEONGGYU LEE

(Paper Review) Abnormal Event Detection in Videos using Generative Adversaria...
(Paper Review) Abnormal Event Detection in Videos using Generative Adversaria...(Paper Review) Abnormal Event Detection in Videos using Generative Adversaria...
(Paper Review) Abnormal Event Detection in Videos using Generative Adversaria...MYEONGGYU LEE
 
Survey of Super Resolution Task (SISR Only)
Survey of Super Resolution Task (SISR Only)Survey of Super Resolution Task (SISR Only)
Survey of Super Resolution Task (SISR Only)MYEONGGYU LEE
 
Simple Review of Single Image Super Resolution Task
Simple Review of Single Image Super Resolution TaskSimple Review of Single Image Super Resolution Task
Simple Review of Single Image Super Resolution TaskMYEONGGYU LEE
 
(Paper Review)Few-Shot Adversarial Learning of Realistic Neural Talking Head ...
(Paper Review)Few-Shot Adversarial Learning of Realistic Neural Talking Head ...(Paper Review)Few-Shot Adversarial Learning of Realistic Neural Talking Head ...
(Paper Review)Few-Shot Adversarial Learning of Realistic Neural Talking Head ...MYEONGGYU LEE
 
(Paper Review)U-GAT-IT: unsupervised generative attentional networks with ada...
(Paper Review)U-GAT-IT: unsupervised generative attentional networks with ada...(Paper Review)U-GAT-IT: unsupervised generative attentional networks with ada...
(Paper Review)U-GAT-IT: unsupervised generative attentional networks with ada...MYEONGGYU LEE
 
(Book Summary) Classification and ensemble(book review)
(Book Summary) Classification and ensemble(book review)(Book Summary) Classification and ensemble(book review)
(Book Summary) Classification and ensemble(book review)MYEONGGYU LEE
 
(Paper Review) Reconstruction of Monte Carlo Image Sequences using a Recurren...
(Paper Review) Reconstruction of Monte Carlo Image Sequences using a Recurren...(Paper Review) Reconstruction of Monte Carlo Image Sequences using a Recurren...
(Paper Review) Reconstruction of Monte Carlo Image Sequences using a Recurren...MYEONGGYU LEE
 
(Paper Review)A versatile learning based 3D temporal tracker - scalable, robu...
(Paper Review)A versatile learning based 3D temporal tracker - scalable, robu...(Paper Review)A versatile learning based 3D temporal tracker - scalable, robu...
(Paper Review)A versatile learning based 3D temporal tracker - scalable, robu...MYEONGGYU LEE
 
(Paper Review)Neural 3D mesh renderer
(Paper Review)Neural 3D mesh renderer(Paper Review)Neural 3D mesh renderer
(Paper Review)Neural 3D mesh rendererMYEONGGYU LEE
 
(Paper Review)3D shape reconstruction from sketches via multi view convolutio...
(Paper Review)3D shape reconstruction from sketches via multi view convolutio...(Paper Review)3D shape reconstruction from sketches via multi view convolutio...
(Paper Review)3D shape reconstruction from sketches via multi view convolutio...MYEONGGYU LEE
 
(Paper Review)Image to image translation with conditional adversarial network...
(Paper Review)Image to image translation with conditional adversarial network...(Paper Review)Image to image translation with conditional adversarial network...
(Paper Review)Image to image translation with conditional adversarial network...MYEONGGYU LEE
 
(Book summary) Ensemble method 2018summerml_study
(Book summary) Ensemble method 2018summerml_study(Book summary) Ensemble method 2018summerml_study
(Book summary) Ensemble method 2018summerml_studyMYEONGGYU LEE
 
(Paper Review)Towards foveated rendering for gaze tracked virtual reality
(Paper Review)Towards foveated rendering for gaze tracked virtual reality(Paper Review)Towards foveated rendering for gaze tracked virtual reality
(Paper Review)Towards foveated rendering for gaze tracked virtual realityMYEONGGYU LEE
 
(Paper Review)Geometrically correct projection-based texture mapping onto a d...
(Paper Review)Geometrically correct projection-based texture mapping onto a d...(Paper Review)Geometrically correct projection-based texture mapping onto a d...
(Paper Review)Geometrically correct projection-based texture mapping onto a d...MYEONGGYU LEE
 
(Papers Review)CNN for sentence classification
(Papers Review)CNN for sentence classification(Papers Review)CNN for sentence classification
(Papers Review)CNN for sentence classificationMYEONGGYU LEE
 
(Paper Review)Kernel predicting-convolutional-networks-for-denoising-monte-ca...
(Paper Review)Kernel predicting-convolutional-networks-for-denoising-monte-ca...(Paper Review)Kernel predicting-convolutional-networks-for-denoising-monte-ca...
(Paper Review)Kernel predicting-convolutional-networks-for-denoising-monte-ca...MYEONGGYU LEE
 

Mais de MYEONGGYU LEE (17)

(Paper Review) Abnormal Event Detection in Videos using Generative Adversaria...
(Paper Review) Abnormal Event Detection in Videos using Generative Adversaria...(Paper Review) Abnormal Event Detection in Videos using Generative Adversaria...
(Paper Review) Abnormal Event Detection in Videos using Generative Adversaria...
 
Survey of Super Resolution Task (SISR Only)
Survey of Super Resolution Task (SISR Only)Survey of Super Resolution Task (SISR Only)
Survey of Super Resolution Task (SISR Only)
 
Simple Review of Single Image Super Resolution Task
Simple Review of Single Image Super Resolution TaskSimple Review of Single Image Super Resolution Task
Simple Review of Single Image Super Resolution Task
 
ICCV 2019 Review
ICCV 2019 ReviewICCV 2019 Review
ICCV 2019 Review
 
(Paper Review)Few-Shot Adversarial Learning of Realistic Neural Talking Head ...
(Paper Review)Few-Shot Adversarial Learning of Realistic Neural Talking Head ...(Paper Review)Few-Shot Adversarial Learning of Realistic Neural Talking Head ...
(Paper Review)Few-Shot Adversarial Learning of Realistic Neural Talking Head ...
 
(Paper Review)U-GAT-IT: unsupervised generative attentional networks with ada...
(Paper Review)U-GAT-IT: unsupervised generative attentional networks with ada...(Paper Review)U-GAT-IT: unsupervised generative attentional networks with ada...
(Paper Review)U-GAT-IT: unsupervised generative attentional networks with ada...
 
(Book Summary) Classification and ensemble(book review)
(Book Summary) Classification and ensemble(book review)(Book Summary) Classification and ensemble(book review)
(Book Summary) Classification and ensemble(book review)
 
(Paper Review) Reconstruction of Monte Carlo Image Sequences using a Recurren...
(Paper Review) Reconstruction of Monte Carlo Image Sequences using a Recurren...(Paper Review) Reconstruction of Monte Carlo Image Sequences using a Recurren...
(Paper Review) Reconstruction of Monte Carlo Image Sequences using a Recurren...
 
(Paper Review)A versatile learning based 3D temporal tracker - scalable, robu...
(Paper Review)A versatile learning based 3D temporal tracker - scalable, robu...(Paper Review)A versatile learning based 3D temporal tracker - scalable, robu...
(Paper Review)A versatile learning based 3D temporal tracker - scalable, robu...
 
(Paper Review)Neural 3D mesh renderer
(Paper Review)Neural 3D mesh renderer(Paper Review)Neural 3D mesh renderer
(Paper Review)Neural 3D mesh renderer
 
(Paper Review)3D shape reconstruction from sketches via multi view convolutio...
(Paper Review)3D shape reconstruction from sketches via multi view convolutio...(Paper Review)3D shape reconstruction from sketches via multi view convolutio...
(Paper Review)3D shape reconstruction from sketches via multi view convolutio...
 
(Paper Review)Image to image translation with conditional adversarial network...
(Paper Review)Image to image translation with conditional adversarial network...(Paper Review)Image to image translation with conditional adversarial network...
(Paper Review)Image to image translation with conditional adversarial network...
 
(Book summary) Ensemble method 2018summerml_study
(Book summary) Ensemble method 2018summerml_study(Book summary) Ensemble method 2018summerml_study
(Book summary) Ensemble method 2018summerml_study
 
(Paper Review)Towards foveated rendering for gaze tracked virtual reality
(Paper Review)Towards foveated rendering for gaze tracked virtual reality(Paper Review)Towards foveated rendering for gaze tracked virtual reality
(Paper Review)Towards foveated rendering for gaze tracked virtual reality
 
(Paper Review)Geometrically correct projection-based texture mapping onto a d...
(Paper Review)Geometrically correct projection-based texture mapping onto a d...(Paper Review)Geometrically correct projection-based texture mapping onto a d...
(Paper Review)Geometrically correct projection-based texture mapping onto a d...
 
(Papers Review)CNN for sentence classification
(Papers Review)CNN for sentence classification(Papers Review)CNN for sentence classification
(Papers Review)CNN for sentence classification
 
(Paper Review)Kernel predicting-convolutional-networks-for-denoising-monte-ca...
(Paper Review)Kernel predicting-convolutional-networks-for-denoising-monte-ca...(Paper Review)Kernel predicting-convolutional-networks-for-denoising-monte-ca...
(Paper Review)Kernel predicting-convolutional-networks-for-denoising-monte-ca...
 

Último

TeamStation AI System Report LATAM IT Salaries 2024
TeamStation AI System Report LATAM IT Salaries 2024TeamStation AI System Report LATAM IT Salaries 2024
TeamStation AI System Report LATAM IT Salaries 2024Lonnie McRorey
 
Take control of your SAP testing with UiPath Test Suite
Take control of your SAP testing with UiPath Test SuiteTake control of your SAP testing with UiPath Test Suite
Take control of your SAP testing with UiPath Test SuiteDianaGray10
 
Gen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfGen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfAddepto
 
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek SchlawackFwdays
 
Scanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsScanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsRizwan Syed
 
DevoxxFR 2024 Reproducible Builds with Apache Maven
DevoxxFR 2024 Reproducible Builds with Apache MavenDevoxxFR 2024 Reproducible Builds with Apache Maven
DevoxxFR 2024 Reproducible Builds with Apache MavenHervé Boutemy
 
A Deep Dive on Passkeys: FIDO Paris Seminar.pptx
A Deep Dive on Passkeys: FIDO Paris Seminar.pptxA Deep Dive on Passkeys: FIDO Paris Seminar.pptx
A Deep Dive on Passkeys: FIDO Paris Seminar.pptxLoriGlavin3
 
The State of Passkeys with FIDO Alliance.pptx
The State of Passkeys with FIDO Alliance.pptxThe State of Passkeys with FIDO Alliance.pptx
The State of Passkeys with FIDO Alliance.pptxLoriGlavin3
 
Dev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio WebDev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio WebUiPathCommunity
 
How to write a Business Continuity Plan
How to write a Business Continuity PlanHow to write a Business Continuity Plan
How to write a Business Continuity PlanDatabarracks
 
What is DBT - The Ultimate Data Build Tool.pdf
What is DBT - The Ultimate Data Build Tool.pdfWhat is DBT - The Ultimate Data Build Tool.pdf
What is DBT - The Ultimate Data Build Tool.pdfMounikaPolabathina
 
The Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptx
The Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptxThe Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptx
The Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptxLoriGlavin3
 
Advanced Computer Architecture – An Introduction
Advanced Computer Architecture – An IntroductionAdvanced Computer Architecture – An Introduction
Advanced Computer Architecture – An IntroductionDilum Bandara
 
Moving Beyond Passwords: FIDO Paris Seminar.pdf
Moving Beyond Passwords: FIDO Paris Seminar.pdfMoving Beyond Passwords: FIDO Paris Seminar.pdf
Moving Beyond Passwords: FIDO Paris Seminar.pdfLoriGlavin3
 
Streamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project SetupStreamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project SetupFlorian Wilhelm
 
SALESFORCE EDUCATION CLOUD | FEXLE SERVICES
SALESFORCE EDUCATION CLOUD | FEXLE SERVICESSALESFORCE EDUCATION CLOUD | FEXLE SERVICES
SALESFORCE EDUCATION CLOUD | FEXLE SERVICESmohitsingh558521
 
Commit 2024 - Secret Management made easy
Commit 2024 - Secret Management made easyCommit 2024 - Secret Management made easy
Commit 2024 - Secret Management made easyAlfredo García Lavilla
 
From Family Reminiscence to Scholarly Archive .
From Family Reminiscence to Scholarly Archive .From Family Reminiscence to Scholarly Archive .
From Family Reminiscence to Scholarly Archive .Alan Dix
 
The Role of FIDO in a Cyber Secure Netherlands: FIDO Paris Seminar.pptx
The Role of FIDO in a Cyber Secure Netherlands: FIDO Paris Seminar.pptxThe Role of FIDO in a Cyber Secure Netherlands: FIDO Paris Seminar.pptx
The Role of FIDO in a Cyber Secure Netherlands: FIDO Paris Seminar.pptxLoriGlavin3
 

Último (20)

TeamStation AI System Report LATAM IT Salaries 2024
TeamStation AI System Report LATAM IT Salaries 2024TeamStation AI System Report LATAM IT Salaries 2024
TeamStation AI System Report LATAM IT Salaries 2024
 
Take control of your SAP testing with UiPath Test Suite
Take control of your SAP testing with UiPath Test SuiteTake control of your SAP testing with UiPath Test Suite
Take control of your SAP testing with UiPath Test Suite
 
DMCC Future of Trade Web3 - Special Edition
DMCC Future of Trade Web3 - Special EditionDMCC Future of Trade Web3 - Special Edition
DMCC Future of Trade Web3 - Special Edition
 
Gen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfGen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdf
 
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
"Subclassing and Composition – A Pythonic Tour of Trade-Offs", Hynek Schlawack
 
Scanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsScanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL Certs
 
DevoxxFR 2024 Reproducible Builds with Apache Maven
DevoxxFR 2024 Reproducible Builds with Apache MavenDevoxxFR 2024 Reproducible Builds with Apache Maven
DevoxxFR 2024 Reproducible Builds with Apache Maven
 
A Deep Dive on Passkeys: FIDO Paris Seminar.pptx
A Deep Dive on Passkeys: FIDO Paris Seminar.pptxA Deep Dive on Passkeys: FIDO Paris Seminar.pptx
A Deep Dive on Passkeys: FIDO Paris Seminar.pptx
 
The State of Passkeys with FIDO Alliance.pptx
The State of Passkeys with FIDO Alliance.pptxThe State of Passkeys with FIDO Alliance.pptx
The State of Passkeys with FIDO Alliance.pptx
 
Dev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio WebDev Dives: Streamline document processing with UiPath Studio Web
Dev Dives: Streamline document processing with UiPath Studio Web
 
How to write a Business Continuity Plan
How to write a Business Continuity PlanHow to write a Business Continuity Plan
How to write a Business Continuity Plan
 
What is DBT - The Ultimate Data Build Tool.pdf
What is DBT - The Ultimate Data Build Tool.pdfWhat is DBT - The Ultimate Data Build Tool.pdf
What is DBT - The Ultimate Data Build Tool.pdf
 
The Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptx
The Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptxThe Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptx
The Fit for Passkeys for Employee and Consumer Sign-ins: FIDO Paris Seminar.pptx
 
Advanced Computer Architecture – An Introduction
Advanced Computer Architecture – An IntroductionAdvanced Computer Architecture – An Introduction
Advanced Computer Architecture – An Introduction
 
Moving Beyond Passwords: FIDO Paris Seminar.pdf
Moving Beyond Passwords: FIDO Paris Seminar.pdfMoving Beyond Passwords: FIDO Paris Seminar.pdf
Moving Beyond Passwords: FIDO Paris Seminar.pdf
 
Streamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project SetupStreamlining Python Development: A Guide to a Modern Project Setup
Streamlining Python Development: A Guide to a Modern Project Setup
 
SALESFORCE EDUCATION CLOUD | FEXLE SERVICES
SALESFORCE EDUCATION CLOUD | FEXLE SERVICESSALESFORCE EDUCATION CLOUD | FEXLE SERVICES
SALESFORCE EDUCATION CLOUD | FEXLE SERVICES
 
Commit 2024 - Secret Management made easy
Commit 2024 - Secret Management made easyCommit 2024 - Secret Management made easy
Commit 2024 - Secret Management made easy
 
From Family Reminiscence to Scholarly Archive .
From Family Reminiscence to Scholarly Archive .From Family Reminiscence to Scholarly Archive .
From Family Reminiscence to Scholarly Archive .
 
The Role of FIDO in a Cyber Secure Netherlands: FIDO Paris Seminar.pptx
The Role of FIDO in a Cyber Secure Netherlands: FIDO Paris Seminar.pptxThe Role of FIDO in a Cyber Secure Netherlands: FIDO Paris Seminar.pptx
The Role of FIDO in a Cyber Secure Netherlands: FIDO Paris Seminar.pptx
 

HDR and Tone Mapping Seminar Guide

  • 1. 응용개발사업부 HDR & Tone Mapping (1/36) 2021/03/19 HDR & Tone Mapping HDR 기술의 소개부터 딥러닝 기반 최신 톤 매핑 기법까지 Tech Seminar Presented by 응용개발사업부 이명규
  • 2. 응용개발사업부 HDR & Tone Mapping (2/36) I N D E X 01 02 03 Introduction HDR & Tone Mapping Featured Papers
  • 3. 응용개발사업부 HDR & Tone Mapping (3/36) Introduction Part 01 1. HDR Image 2. HDR In Various Fields
  • 4. 응용개발사업부 HDR & Tone Mapping (4/36) HDR Image 1-1 • HDR: “Content with a Wider Range of Brightness and Color” • HDR을 잘 표현하기 위해서는 새로운 디스플레이 장치가 필요 • 기존 콘텐츠: 최대 100nit 밝기, 709 Gamut* • HDR 콘텐츠: 최대 10,000nit 밝기, 2020 Gamut HDR LDR *Gammut: 색상과 채도 (밝기와 대비를 의미하는 Gamma와 구분) High-Dynamic Range (HDR) Demystified Object Approx. Luminance (nits) Sun 1,600,000,000 Arc Lamp 150,000,000 Maximum Visual Tolerance 50,000 Cloud (Sunny Day) 35,000 2016 UHD TV 800~1,000 Typical Computer Screen 100~300 White Paper Under the Lamp 50 Night Sky 0.001 Threshold of Vision 0.000003
  • 5. 응용개발사업부 HDR & Tone Mapping (5/36) HDR In Various Fields 1-2 VOD Service Gaming Photography High Dynamic Range Imaging 기술 및 최근 동향 (The Korean Information Display Society, 2019) • HDR Imaging? • 입력 영상에 대해 노출 정도를 다르게 설정 후 취득한 여러 이미지를 합성함으로써 Dynamic Range 증강 • 카메라에 의해 취득된 HDR 영상을 LCD, OLED 등의 디스플레이에서 정확히 표현할 수 있도록 하는 것
  • 6. 응용개발사업부 HDR & Tone Mapping (6/36) HDR In Various Fields 1-2 High Dynamic Range Imaging Technology
  • 7. 응용개발사업부 HDR & Tone Mapping (7/36) HDR & Tone Mapping Part 02 1. What is HDR? 2. What is Tone Mapping?
  • 8. 응용개발사업부 HDR & Tone Mapping (8/36) ↳ What is HDR? 2-1 Dynamic Range of Human Eye Use Cyclone® V SoC FPGA to Create Real-time HDR Video(Intel), Investigation on the Use of HDR Images for Cultural Heritage Documentation • Contrast Ratio란? • 디스플레이가 출력 가능한 가장 밝은 색(white)과 어두운 색(Black)의 휘도(Luminance) 명암 비율
  • 9. 응용개발사업부 HDR & Tone Mapping (9/36) ↳ What is HDR? 2-1 Dynamic Range of Human Eye 'Light Adaptation'(Vision Models for High Dynamic Range and Wide Colour Gamut Imaging, 2020), “High Quality High Dynamic Range Imaging” Cone Function (추상체 수용기 감지범위) No Moon Moonlight (Fullmoon) Early Twilight Store, Office Outdoors (Sunny) Starlight 𝟏𝟎−𝟔 𝟏𝟎−𝟓 𝟏𝟎−𝟒 𝟏𝟎−𝟑 𝟏𝟎−𝟐 𝟏𝟎−𝟏 𝟏 𝟏𝟎𝟏 𝟏𝟎𝟐 𝟏𝟎𝟑 𝟏𝟎𝟒 𝟏𝟎𝟓 𝟏𝟎𝟔 Luminance (𝒄𝒅/𝒎𝟐 ) Rod Function (간상체 수용기 감지범위) HDR Display Normal Display Image Capturing (Loss Dynamic Range) Domain of Human Vision: ≈ 𝟏𝟎−𝟒 ~ ≈ 𝟏𝟎𝟔 Conventional Image: −𝟏 ~ ≈ 𝟏𝟎𝟐 “LDR Image” (8-bit Integer [0~255]) “HDR Image” (32-bit Floating Point) HDR Imaging (Recover lost Dynamic Range)
  • 10. 응용개발사업부 HDR & Tone Mapping (10/36) ↳ What is HDR? 2-1 LDR VS. HDR High Dynamic Range Imaging 기술 및 최근 동향 • Contrast Ratio에 따른 신호 구분 • 명암비가 1000:1보다 작은 경우: LDR(Low Dynamic Range) 또는 SDR(Standard Dynamic Range) • 명암비가 1000:1~100,000:1인 경우: EDR (Enhanced Dynamic Range) • 명암비가 100,000:1보다 큰 경우: HDR(High Dynamic Range) • HDR 영상의 활용을 위해서는 디스플레이의 영상 신호처리 기술과 함께 색 재현율에 대한 기본적인 이해 필요 Today’s Topic
  • 11. 응용개발사업부 HDR & Tone Mapping (11/36) ↳ What is HDR? 2-1 LDR VS. HDR [0107 박민근] 쉽게 배우는 hdr과 톤맵핑 • LDR(SDR) ➢0~255로 256단계 색상 표현 가능 (𝟐𝟓𝟔𝟑 = 𝟏𝟔, 𝟕𝟕𝟕, 𝟐𝟏𝟔개 색 표현 가능) ➢픽셀당 4Byte씩 총 32-bit 사용 (R: 8-bit, G: 8-bit, B: 8-bit, A: 8-bit) • HDR(OpenEXR format) ➢색상 당 16-bit Floating Point로 표현 ➢채널 당 부호 1-bit, 가수에 10-bit, 지수에 5-bit 사용 (알파채널 포함 총 64-bit 버퍼 사용) ➢약 𝟒. 𝟒 × 𝟏𝟎𝟏𝟐 = 4.4e+12 색상 표현 가능 (≈ 𝟒, 𝟒𝟎𝟎, 𝟎𝟎𝟎, 𝟎𝟎𝟎, 𝟎𝟎𝟎개)
  • 12. 응용개발사업부 HDR & Tone Mapping (12/36) ↳ What is HDR? 2-1 LDR VS. HDR High-Dynamic Range (HDR) Demystified, [디스플레이 톺아보기] ㉚ HDR(High Dynamic Range)의 이해 *SDR: Standard Dynamic Range (30 nit laptops in low power mode, 600 nit HDTVs in vivid mode) Real World Mastered for SDR* Mastered for HDR
  • 13. 응용개발사업부 HDR & Tone Mapping (13/36) ↳ What is HDR? 2-1 Color Bit [디스플레이 톺아보기] ㉙ 디스플레이 색심도(Color Depth)의 이해, Converting Color Depth | Color 8-bit 3-bit 8-bit 24-bit 10-bit
  • 14. 응용개발사업부 HDR & Tone Mapping (14/36) ↳ What is HDR? 2-1 HDR Standard Formats [디스플레이 톺아보기] ㉚ HDR(High Dynamic Range)의 이해 HDR10 Dolby Vision Color Gamut (색역) Color Depth (색심도) 10bit 12bit Peak Luminance (최대밝기) 1000nits 10000nits Meta Data Static (콘텐츠별 설정) Dynamic (프레임별 설정) BT.2020
  • 15. 응용개발사업부 HDR & Tone Mapping (15/36) ↳ What is HDR? 2-1 HDR Standard Formats High-dynamic-range video HDR10 HDR10+ Dolby Vision HLG10 CTA Samsung Dolby NHK and BBC 2015 2017 2014 2015 Free Free (for content company), Yearly license (for manufacturer) Proprietary Free Static (SMPTE ST 2086, MaxFALL, MaxCLL) Dynamic Dynamic (Dolby Vision L0, L1, L2 trim, L8 trim) None PQ PQ PQ (Not always) HLG 10 bit 10 bit (or more) 10 bit or 12 bit 10 bit Technical limit 10,000 nits 10,000 nits 10,000 nits Variable Contents No rules 1,000 - 4,000 nits (common) No rules 1,000 - 4,000 nits (common) (At least 1,000 nits) 4,000 nits common 1,000 nits common Technical limit Rec. 2020 Rec. 2020 Rec. 2020 Rec. 2020 Contents DCI-P3 (common) DCI-P3 (common) At least DCI-P3 DCI-P3 (common) None HDR10 It depends on the profile used: - No compatibility - SDR - HDR10 - HLG - Ultra HD Blu-Ray - UHD-TV (Rec.2020) - SDR (Rec.709) with color distortion Transfer function Bit Depth Peak  luminance Color primaries Backward compatibility Metadata Developed by Year Cost Technical charasteristics
  • 16. 응용개발사업부 HDR & Tone Mapping (16/36) ↳ What is HDR? 2-1 모니터의 영상 신호 처리 High Dynamic Range Imaging 기술 및 최근 동향 • 디스플레이는 인간의 시각적 특성을 고려해 신호 처리 • 베버의 법칙에 따라, 정해진 최대 표현량을 정확히 고려해 최적의 화질을 재생할 수 있는 비선형 관계 이해가 중요 • 단순히 선형적으로 빛의 밝기를 표현하면 Posterization 현상이 발생하며, 이를 해결하기 위해 감마 보정 기법 사용 • HDR Imaging은 기존 디스플레이보다 높은 10~12-bit 범위에서 영상을 표현하므로, 주어진 bit depth 내에서 입력/출력 신호 간 관계를 적절하게 정의해야 함 • HDR Imaging은 또한 기존에 사용되던 Standard RGB 대비 170% 증가된 색역(e.g. BT Rec.2020)을 지원해야 함.
  • 17. 응용개발사업부 HDR & Tone Mapping (17/36) ↳ • “인간의 감각은 선형적이지 않다” ➢초기에 받은 자극의 정도에 따라 나중에 받는 자극의 수용폭이 달라진다는 생리학 이론 Weber’s Law 손바닥에 100g의 무게부터 조금씩 무게를 늘려 나갔을 때 102g에서 최초로 무게가 다르다는 것을 느낄 수가 있었고, 200g의 물건을 올려놓았을 때는 204g에서 최초로 무게가 다르다는 것을 느낄 수 있었다. 이 실험에서 베버는 최초로 차이를 느낄 수 있을 때의 자극의 증가량 2g, 4g(R 절대판별역)과 처음 올려놓은 표준자극 100g과 200g(R)의 비(R/R상대판별역)는 항상 비례적으로 일정하다는 것을 발견했다. 이렇게 감각으로 구별할 수 있는 한계는 물리적인양의 차이가 아니고, 그 비율관계에 의해 결정된다는 것이다. 베버-페흐너의 법칙(Weber-Fechner’s law), UX디자이너가 알아야 할 심리학 법칙 5가지 What is HDR? 2-1
  • 18. 응용개발사업부 HDR & Tone Mapping (18/36) ↳ Gamma Correction 머신 비전 ISP – 17. Gammar Corection, 감마보정 Gamma Correction • “모니터는 실제보다 어둡게 빛을 표현한다!” • 사람의 눈은 어두운 환경에서 작은 밝기의 변화에도 민감하게 반응 (반면 밝은 환경에서는 작은 밝기의 변화에 둔감하게 반응) • “인간의 시각은 비선형적이므로, 굳이 인간이 잘 느끼지 못하는 부분까지 정밀하게 계산할 필요는 없다” 𝒇 𝒙 = 𝑮𝒂𝒊𝒏 × 𝑿𝒈𝒂𝒎𝒎𝒂 + 𝒐𝒇𝒇𝒔𝒆𝒕 Nonlinear Transfer Function Intensity (HDR) Transferred Intensity (Display) 일반적인 CRT 모니터의 감마 What is HDR? 2-1
  • 19. 응용개발사업부 HDR & Tone Mapping (19/36) ↳ Gamma Correction Lighting Shading by John Hable, 이미지 파일 감마 (Image File Gamma)와 디스플레이 감마 (Display Gamma) • 이미지는 일반적으로 1/2.2로 감마를 적용한 상태에서 저장 • 디스플레이 출력 시 올바른 색상 표현을 위해 저장 시의 Gamma를 상쇄해야 함 • 올바른 HDR 결과를 얻기 위해서는 감마 보정 작업을 거쳐야 한다! What is HDR? 2-1
  • 20. 응용개발사업부 HDR & Tone Mapping (20/36) ↳ What is Tone Mapping? 2-2 • 따라서 HDR 포맷을 디스플레이 출력이 가능한 휘도 범위로 변환 처리해야 함. • 톤 매핑은 HDR 범위를 인간 색 인지에 기반해 Display에서 지원하는 범위로 맞추는 작업 • Tone Mapping 기법의 종류 • HDR to HDR • HDR to LDR • LDR to HDR (“Inverse Tone Mapping”) Tone Mapping • HDR(FP16) ➢ 색상 당 16-bit Floating Point로 표현 ➢ 채널 당 부호 1-bit, 가수에 10-bit, 지수에 5-bit 사용 (알파채널 포함 총 64-bit 버퍼 사용) ➢ 약 𝟒. 𝟒 × 𝟏𝟎𝟏𝟐 = 4.4e+12 색상 표현 가능 (≈ 𝟒, 𝟒𝟎𝟎, 𝟎𝟎𝟎, 𝟎𝟎𝟎, 𝟎𝟎𝟎) “아무리 실수 범위에서 넓게 계산해도 일반 모니터는 4Byte RGB만 출력 가능” Today’s Topic
  • 21. 응용개발사업부 HDR & Tone Mapping (21/36) ↳ What is Tone Mapping? 2-2 Tone Mapping SDR Content SDR Display SDR Experience SDR Content HDR Display SDR Experience HDR Content HDR Display HDR Experience HDR Content SDR Display Bad Experience HDR Content HDR▶SDR Tone Mapping SDR Display SDR Experience High-Dynamic Range (HDR) Demystified
  • 22. 응용개발사업부 HDR & Tone Mapping (22/36) ↳ What is Tone Mapping? 2-2 • HDR Imaging에는 비싼 장비와 상당한 연산 시간이 필요 ➢ 따라서 LDR 이미지만으로 Real World Luminance를 추정하는 역 톤 매핑 기법의 필요성 대두 • 톤 매핑 과정에서 Linear RGB를 CIE 1931 XYZ Color Space로 변환 ➢ 𝑿 = 𝟎. 𝟒𝟏𝟐𝑹𝒘 + 𝟎. 𝟑𝟓𝟕𝑮𝒘 + 𝟎. 𝟏𝟖𝑩𝒘 𝒀 = 𝑳𝑾 = 𝟎. 𝟐𝟏𝟑𝑹𝒘 + 𝟎. 𝟕𝟏𝟓𝑮𝒘 + 𝟎. 𝟎𝟕𝟐𝑩𝒘, where 𝐿𝑊 is real world luminance(≈HDR) 𝒁 = 𝟎. 𝟎𝟏𝟗𝑹𝒘 + 𝟎. 𝟏𝟏𝟗𝑮𝒘 + 𝟎. 𝟗𝟓𝑩𝒘 ➢ XYZ Color Space는 인간의 색채 인지 연구를 바탕으로 만들어진 색 공간 (따라서 다른 색공간의 기본이 됨) • 톤 매핑 과정에서 압축된 정보의 손실이 발생하므로 역 톤 매핑은 ill-posed problem에 해당 • 역 톤 매핑은 𝑹𝒅, 𝑮𝒅, 𝑩𝒅만 주어지므로 Real World Color(𝑹𝒘, 𝑮𝒘, 𝑩𝒘)로 돌아가는 변환인 𝑳𝑾을 알 수 없음 Inverse Tone Mapping "Inverse tone mapping“, ITU-R Recommendation BT.709 RGB, CIE 1931 color space, high Dynamic Range Imaging: Acquisition, Display and Image-Based Lighting 𝑹𝒅 𝑮𝒅 𝑩𝒅 = 𝑳𝒅 𝑹𝒘 𝑳𝒘 𝑳𝒅 𝑮𝒘 𝑳𝒘 𝑳𝒅 𝑩𝒘 𝑳𝒘 ▲ Tone Mapping 𝑹𝒘 𝑮𝒘 𝑩𝒘 = 𝑳𝒘 𝑹𝒅 𝑳𝒅 𝑳𝒘 𝑮𝒅 𝑳𝒅 𝑳𝒘 𝑩𝒅 𝑳𝒅 ▲ Inverse Tone Mapping Compressed Colors World Colors ???
  • 23. 응용개발사업부 HDR & Tone Mapping (23/36) ↳ What is Tone Mapping? 2-2 • Previous Approaches ➢“The Reproduction of Specular Highlights on High Dynamic Range Displays” ✓ 출력 디스플레이 특성을 고려해 Dynamic Range를 적응적으로 조절하는 방법 ✓ 빛이 방출되거나 반사되는 영역에 대해 더 많은 정보를 부여해 표현력 향상 ➢“Ldr2Hdr: on-the-fly reverse tone mapping of legacy video and photographs” ✓ 입력 영상 내에서 상위 계조를 갖는 영역 검출 후 더 많은 정보를 부여해 표현력 향상 ✓ 빛 분포 고려 후 픽셀 표현 가능 범위를 넘어 saturation된 상위 계조에 집중 ✓“Physiological inverse tone mapping based on retina response” ✓ 인간의 시각적 특성을 반영한 Perceptual Brightness를 정의해 입출력 영상 신호 간 상관 관계를 적응적으로 정의 Inverse Tone Mapping: Method High Dynamic Range Imaging 기술 및 최근 동향 (The Korean Information Display Society, 2019)
  • 24. 응용개발사업부 HDR & Tone Mapping (24/36) ↳ What is Tone Mapping? 2-2 • Deep-Learning Based Approaches ➢“HDR image reconstruction from a single exposure using deep CNNs” ✓ 입력 영상에 대해 상위/하위 계조를 나눈 후 하위 계조는 원본 신호를, 상위 계조는 CNN을 통해 추론된 신호를 사용(단, 전체적으로 어두운 영상에서만 잘 작동) ➢“Deep Recursive HDRI: Inverse Tone Mapping using Generative Adversarial Networks” ✓ GAN을 활용해 성능을 향상(L1-Loss, Adversarial Loss, U-Net, PatchGAN) Inverse Tone Mapping: Method High Dynamic Range Imaging 기술 및 최근 동향 (The Korean Information Display Society, 2019)
  • 25. 응용개발사업부 HDR & Tone Mapping (25/36) Featured Papers Part 03 1. HDRCNN (SIGGRAPH ASIA 2017) 2. SingleHDR (CVPR 2020) 3. LDR2HDR (SIGGRAPH 2020)
  • 26. 응용개발사업부 HDR & Tone Mapping (26/36) ↳ HDRCNN 3-1 "HDR image reconstruction from a single exposure using deep CNNs" (SIGGRAPH ASIA 2017, GABRIEL EILERTSEN et al.) [Paper, Code] https://arxiv.org/abs/1710.07480 “Highlight Information을 유지하면서(초록 박스) Saturate된 색상 영역의 Color Detail 복원”
  • 27. 응용개발사업부 HDR & Tone Mapping (27/36) ↳ Convert to logarithmic HDR Domain Performs Bilinear Up-sampling Performs Addition Predicted HDR image in log domain HDRCNN 3-1 Network Structure https://arxiv.org/abs/1710.07480
  • 28. 응용개발사업부 HDR & Tone Mapping (28/36) ↳ HDRCNN 3-1 Problem Formulation & Loss Function https://arxiv.org/abs/1710.07480, Modeling the Space of Camera Response Functions 𝓛 ෝ 𝒚, 𝑯 = 𝟏 𝟑𝑵 ෍ 𝒊,𝒄 𝜶𝒊(ෝ 𝒚𝒊,𝒄 − 𝐥𝐨𝐠(𝑯𝒊,𝒄 + 𝝐)) 𝟐 𝑯𝒊,𝒄 ∈ ℝ+ # of Pixels Linear GT HDR Predicted Log HDR ▲ 201 Real-world Response Functions Database(“DoRF”) ෡ 𝑯𝒊,𝒄 = 𝟏 − 𝜶𝒊 𝒇−𝟏 𝑫𝒊,𝒄 + 𝜶𝒊𝐞𝐱𝐩(ෝ 𝒚𝒊, 𝒄) Final Reconstructed HDR pixels (where 𝒊=spatial index, 𝒄=channel) Input LDR Pixel Predicted output (log domain) Inverse Camera Function (transform input to linear domain) Blending Factor 𝜶𝒊 = 𝐦𝐚𝐱(𝟎,𝒎𝒂𝒙𝒄 𝑫𝒊,𝒄 −𝝉) 𝟏−𝝉 , where 𝝉 =0.95 (하이라이트 부근의 Banding artifact를 방지하기 위한 factor)
  • 29. 응용개발사업부 HDR & Tone Mapping (29/36) ↳ SingleHDR 3-2 "Single-Image HDR Reconstruction by Learning to Reverse the Camera Pipeline" (CVPR 2020, Yu-Lun Liu et al.) [Paper, Code] https://arxiv.org/abs/2004.01179 “카메라 파이프라인을 역으로 모델링해 학습함으로써 소실된 디테일 정보 복원”
  • 30. 응용개발사업부 HDR & Tone Mapping (30/36) ↳ SingleHDR 3-2 Network Structure https://arxiv.org/abs/2004.01179 ▲ LDR Image formation pipeline ▲ Proposed Method
  • 31. 응용개발사업부 HDR & Tone Mapping (31/36) ↳ SingleHDR 3-2 Network Structure https://arxiv.org/abs/2004.01179 ▲ Linearization Network ▲ Hallucination Network
  • 32. 응용개발사업부 HDR & Tone Mapping (32/36) ↳ LDR2HDR 3-3 "Single Image HDR Reconstruction Using a CNN with Masked Features and Perceptual Loss" (SIGGRAPH 2020, MARCEL SANTANA SANTOS et al.) [Paper, Code] https://people.engr.tamu.edu/nimak/Data/SIGGRAPH20_HDR.pdf “높은 노출로 인해 밝게 타 버린 텍스쳐 영역도 잘 복원”
  • 33. 응용개발사업부 HDR & Tone Mapping (33/36) ↳ LDR2HDR 3-3 Demo Images https://people.engr.tamu.edu/nimak/Data/SIGGRAPH20_HDR.pdf
  • 34. 응용개발사업부 HDR & Tone Mapping (34/36) ↳ LDR2HDR 3-3 Network Structure https://people.engr.tamu.edu/nimak/Data/SIGGRAPH20_HDR.pdf 일반적인 문제와는 달리 타 버린(saturated areas) LDR 영역에 대해서도 잘 작동시키기 위해 Feature Masking 기법을 제안
  • 35. 응용개발사업부 HDR & Tone Mapping (35/36) ↳ LDR2HDR 3-3 Network Structure: Feature Masking https://people.engr.tamu.edu/nimak/Data/SIGGRAPH20_HDR.pdf 𝒁𝒍 = 𝑿𝒍⨀𝑴𝒍 𝑿𝒍+𝟏 = 𝝓𝒍(𝑾𝒍 ∗ 𝒁𝒍 + 𝒃𝒍) 𝑿𝒍 ∈ ℝ𝑯×𝑾×𝑪 , 𝑴𝒍 ∈ 𝟎, 𝟏 𝑯×𝑾×𝑪 𝑊𝑙 and 𝑏𝑙 is weight and bias of the current layer
  • 36. 응용개발사업부 HDR & Tone Mapping (36/36) ↳ LDR2HDR 3-3 Loss Function https://people.engr.tamu.edu/nimak/Data/SIGGRAPH20_HDR.pdf 𝑳 = 𝝀𝟏𝑳𝒓 + 𝝀𝟐𝑳𝒑, 𝒘𝒉𝒆𝒓𝒆 𝝀𝟏 = 𝟔. 𝟎 𝒂𝒏𝒅 𝝀𝟐 = 𝟏. 𝟎 Reconstruction Loss Perception Loss 𝑳𝒓 = (𝟏 − 𝑴 ⊙ (෡ 𝒀 − 𝐥𝐨𝐠(𝑯 + 𝟏)) 𝟏 Loss가 Saturate된 영역에서 계산되도록 마스킹 𝑳𝒑 = 𝝀𝟑𝑳𝝊 + 𝝀𝟒𝑳𝒔 𝑳𝝊 = ෍ 𝒍 𝝓𝒍 𝓣 ෩ 𝑯 − 𝝓𝒍(𝓣(𝑯)) 𝟏 𝑳𝑺 = ෍ 𝒍 𝑮𝒍 𝓣 ෩ 𝑯 − 𝑮𝒍(𝓣(𝑯)) 𝟏 𝑮𝒍 𝑿 = 𝟏 𝑲𝒍 𝝓𝒍 𝑿 𝑻𝝓𝒍 𝑿 . (𝑮𝒍 is Gram matrix of the feature layer 𝒍.)
  • 37. 응용개발사업부 HDR & Tone Mapping (37/36) Thank you for Watching. brstar96@espresomedia.com (이명규)