Deep Learning in Image and Face Intrinsics

Similar documents
Arnold W.M Smeulders Theo Gevers. University of Amsterdam smeulders}

Supplementary Material: Specular Highlight Removal in Facial Images

Introduction to Computer Vision. Week 8, Fall 2010 Instructor: Prof. Ko Nishino

HOW USEFUL ARE COLOUR INVARIANTS FOR IMAGE RETRIEVAL?

Other approaches to obtaining 3D structure

Color and Shading. Color. Shapiro and Stockman, Chapter 6. Color and Machine Vision. Color and Perception

Intrinsic3D: High-Quality 3D Reconstruction by Joint Appearance and Geometry Optimization with Spatially-Varying Lighting

Radiance. Pixels measure radiance. This pixel Measures radiance along this ray

Chapter 7. Conclusions and Future Work

Illumination and Shading

Color Constancy from Hyper-Spectral Data

Active Stereo Vision. COMP 4900D Winter 2012 Gerhard Roth

3D Shape Analysis with Multi-view Convolutional Networks. Evangelos Kalogerakis

Light, Color, and Surface Reflectance. Shida Beigpour

Deep Learning for Virtual Shopping. Dr. Jürgen Sturm Group Leader RGB-D

Computational Imaging for Self-Driving Vehicles

Physics-based Vision: an Introduction

Digital Vision Face recognition

MERGING POINT CLOUDS FROM MULTIPLE KINECTS. Nishant Rai 13th July, 2016 CARIS Lab University of British Columbia

Model-based Enhancement of Lighting Conditions in Image Sequences

Lecture 24: More on Reflectance CAP 5415

3D Scanning. Qixing Huang Feb. 9 th Slide Credit: Yasutaka Furukawa

L2 Data Acquisition. Mechanical measurement (CMM) Structured light Range images Shape from shading Other methods

Machine vision. Summary # 11: Stereo vision and epipolar geometry. u l = λx. v l = λy

Volume Illumination. Visualisation Lecture 11. Taku Komura. Institute for Perception, Action & Behaviour School of Informatics

Lecture 15: Shading-I. CITS3003 Graphics & Animation

Color Space Invariance for Various Edge Types in Simple Images. Geoffrey Hollinger and Dr. Bruce Maxwell Swarthmore College Summer 2003

Face Recognition Markus Storer, 2007

Accurate 3D Face and Body Modeling from a Single Fixed Kinect

3D Photography: Active Ranging, Structured Light, ICP

3D Shape Segmentation with Projective Convolutional Networks

Advanced Vision Practical

Photo-realistic Renderings for Machines Seong-heum Kim

Specular Reflection Separation using Dark Channel Prior

CONTENTS. High-Accuracy Stereo Depth Maps Using Structured Light. Yeojin Yoon

Why study Computer Vision?

MoFA: Model-based Deep Convolutional Face Autoencoder for Unsupervised Monocular Reconstruction

5/14/2018. Experiences with AR and AI in development projects. Agenda. By: Sune Wolff, CTO & Partner at Unity Studios.

Starting this chapter

Light source estimation using feature points from specular highlights and cast shadows

Using a Raster Display Device for Photometric Stereo

arxiv: v2 [cs.cv] 3 Apr 2018

Augmented Reality. Sung-eui Yoon

Game Programming. Bing-Yu Chen National Taiwan University

The Kinect Sensor. Luís Carriço FCUL 2014/15

A Novel Algorithm for Color Image matching using Wavelet-SIFT

Predicting Depth, Surface Normals and Semantic Labels with a Common Multi-Scale Convolutional Architecture David Eigen, Rob Fergus

Multimedia Technology CHAPTER 4. Video and Animation

DECOMPOSING and editing the illumination of a photograph

Non-line-of-sight imaging

Computer Vision: Making machines see

Colored Point Cloud Registration Revisited Supplementary Material

We want to put a CG object in this room

Image Processing Pipeline for Facial Expression Recognition under Variable Lighting

Recap: Features and filters. Recap: Grouping & fitting. Now: Multiple views 10/29/2008. Epipolar geometry & stereo vision. Why multiple views?

General Principles of 3D Image Analysis

CS5670: Computer Vision

The main problem of photogrammetry

How do we draw a picture?

3D Photography: Stereo

Computer Graphics. Bing-Yu Chen National Taiwan University The University of Tokyo

Volume Illumination & Vector Field Visualisation

Depth Sensors Kinect V2 A. Fornaser

DeepSketch2Face: A Deep Learning Based Sketching System for 3D Face and Caricature Modeling. Part I: User Studies on the Interface

ENGN 2911 I: 3D Photography and Geometry Processing Assignment 2: Structured Light for 3D Scanning

Other Reconstruction Techniques

CS4670/5760: Computer Vision Kavita Bala Scott Wehrwein. Lecture 23: Photometric Stereo

Conversion of 2D Image into 3D and Face Recognition Based Attendance System

Computational Foundations of Cognitive Science

Illumination-Robust Face Recognition based on Gabor Feature Face Intrinsic Identity PCA Model

Removing Shadows from Images

Relighting for an Arbitrary Shape Object Under Unknown Illumination Environment

Colour Reading: Chapter 6. Black body radiators

Multiple View Geometry

An ICA based Approach for Complex Color Scene Text Binarization

Assignment #2. (Due date: 11/6/2012)

Depth. Common Classification Tasks. Example: AlexNet. Another Example: Inception. Another Example: Inception. Depth

Epipolar geometry contd.

Image Based Lighting with Near Light Sources

Image Based Lighting with Near Light Sources

High-Fidelity Facial and Speech Animation for VR HMDs

Robot vision review. Martin Jagersand

HISTOGRAMS OF ORIENTATIO N GRADIENTS

Computer Graphics and Visualization. What is computer graphics?

Neue Verfahren der Bildverarbeitung auch zur Erfassung von Schäden in Abwasserkanälen?

Segmentation and Tracking of Partial Planar Templates

3D object recognition used by team robotto

VISOR vision sensors and the Eyesight vision systems

Lighting and Shading

Capturing, Modeling, Rendering 3D Structures

Stereo. 11/02/2012 CS129, Brown James Hays. Slides by Kristen Grauman

Computational Photography and Video: Intrinsic Images. Prof. Marc Pollefeys Dr. Gabriel Brostow

Computer Graphics (CS 4731) Lecture 16: Lighting, Shading and Materials (Part 1)

Deep Tracking: Biologically Inspired Tracking with Deep Convolutional Networks

A NEW ILLUMINATION INVARIANT FEATURE BASED ON FREAK DESCRIPTOR IN RGB COLOR SPACE

High-Fidelity Augmented Reality Interactions Hrvoje Benko Researcher, MSR Redmond

Deep Learning: Image Registration. Steven Chen and Ty Nguyen

Waleed Pervaiz CSE 352

Illumination and Shading

Multi-sensor Gaze-tracking

Transcription:

Deep Learning in Image and Face Intrinsics Theo Gevers Professor in Computer Vision: University of Amsterdam Co-founder: Sightcorp 3DUniversum Scanm WIC, Eindhoven, February 24th, 2018

Amsterdam A.I. - Computer Vision Group: from the lab to the real world 1. Image Intrinsics 2. Face Intrinsics 3. 3D Reconstruction Intrinsics

Intrinsic images: Old School Approach

Reflection Model C m b n, s e f d m n, s, v C s e c f C d e n s v f C c surface albedo illumination object surface normal illumination direction viewer s direction sensor sensitivity specular scene & viewpoint invariant scene dependent object shape variant scene dependent viewpoint variant camera dependent surface dependent

For giving the red, green and blue sensor response under white light. Consider normalized color: Then: Also holds for g and b. In the sequel: Consider the body reflection term: rgb Photometric Invariance: Proof d f, em C C b WB s n B G R B b B G R G g B G R R r,, d f d f d f, em d f, em r B G R b R b n s n s WB WB WB WB B G R C,, d f k C C

Colour Invariance / rgb space Original RGB image 3D plot of RGB image rgb image 3D plot of rgb image

Colour Invariance - Summary shadows shading highlights ill. intensity ill. Colour I - - - - - R,G,B - - - - - r,g,b + + - + - c1,c2,c3 + + - + - Hue + + + + - l1,l2,l3 + + + + - m1m2m3 + + - + + - no invariance + invariance

Summary: Intrinsic Images Old School Reflection models help to understand the image formation process. Color invariance at the pixel. White light source or successful color constancy. Be aware of instabilities d f c e, m d f e, m C C s C b, v n s n s

Intrinsic images: New School Approach AT THE PIXEL

ConvNet: Encoder-Decoder Style Model Decoders Encoder S I R

Intrinsic images: New School End-to-end solution but no insight in why and how it works. Still outperformed by old school methods in terms of accuracy. Simple loss functions based on reconstruction loss ignoring the image formation model.

ConvNet: Encoder-Decoder Style

ShapeNet Rendered Dataset

Qualitative Results: MIT Dataset

IntrinsicNet: Conclusion Advantage + Integrated image formation model + Can capture most of the color information & shading + Colors are more vivid and eliminates most of the color artifacts + Fast and more memory efficient To be improved - Edge sharpness Next Retinex revisited

Intrinsic images: Old School Approach RETINEX DERIVATIVES

C C f And narrow-band filters: Then: Assuming body reflection Therefore, we have: Von Kries Model B G R C d f e m C C b,, for, s n B G R C e, m C C C b,, for n s R R b e, m R n s B G b e, m G n s B B b e, m B n s

Solving the Inverse Problem I Image e ˆ d 2 Laplacian I 2 eˆ 2 t Thresholding T d 2 Lightness l x, y Find lightness lx,y from tx,y: Poisson s Equation x 2 2 2 l t 2 lx, y tx, y 2 y We have to find gx,y which satisfies x, y tu, vgx u y v l, x, y tx, y gx y l, dudv

Intrinsic Images Original Re-integrated Colour Channels Edge Maps of Channels Shadow Edges Removed

Intrinsic images: New School Approach Retinex-inspired ConvNets

RetiNet: Retinex-Inspired ConvNet RGB Differentiation R RGB Integration x, y gx y t, RGB d 2 eˆ 2 S RGB t x, y gx, y d R S

ShapeNet Rendered Dataset

Qualitative Results: MIT Dataset

Summary Intrinsic image decomposition End-to-end solution. Full testing on intrinsic image decomposition using synthetic and real data. Full model including light source and highlights. Raw data preprocessing and augmentation.

Amsterdam A.I. - Computer Vision Group: from the lab to the real world 1. Image Intrinsics 2. Face Intrinsics 3. 3D Reconstruction Intrinsics

Basic Emotions

Face Analysis by Deep Learning Frame Happy WWW.SIGHTCORP.COM

Sightcorp WWW.SIGHTCORP.COM

FaceWarehouse: a 3D Facial Expression Database for Visual Computing Chen Cao Yanlin Weng Shun Zhou Yiying Tong Kun Zhou State Key Lab of CAD&CG, Zhejiang University Michigan State University

Face Analysis: The Face Model

Face Analysis: The Camera Model

Face Analysis

Co nv 16 Co nv 32 Co nv 64 Co nv 128 Co nv 256 Co nv 256 DeCo nv 256 DeConv 256 DeConv 128 DeConv 64 DeConv 32 DeConv 16 DeCo nv 256 DeConv 256 DeConv 128 DeConv 64 DeConv 32 DeConv 16 DeCo nv 256 DeConv 256 DeConv 128 DeConv 64 DeConv 32 DeConv 16 Decoder- Shading Decoder -Specular Decoder- Albedo Encoder CNN Model: Supervised

Illumination Changes IMAGE ALBEDO SHADING SPECULARITY

Illumination, Geometry, Albedo Changes IMAGE ALBEDO SHADING SPECULARITY

CNN Model: Multi-Task Unsupervised Encoder Decoder rendering

Consumer Applications

Question: Is your presidency fake?

A.I. Face Technology Powered by:

Amsterdam A.I. - Computer Vision Group: from the lab to the real world 1. Image Intrinsics 2. Face Intrinsics 3. 3D Reconstruction Intrinsics

The True Depth camera of the iphone X: Face ID Apple s new TrueDepth camera for the iphone X sure works by using a projector to cast 30,000 dots on your face, which it then reads with an infrared camera like the Microsoft Kinect.

Consumer Applications: AR and VR Hololens Lenovo phab 2 pro

Retail & Fashion: 3D scans allow proper sizing for online shopping no more wasting time and product returns! Scan Faces and Bodies using Mobile Devices Virtual fitting, recommendation and tailoring 3DU has developed deep learning based automatic recognition algorithms which automatically detects 50K facial points eyes, temples, nose bridge, pupils, mouth corners etc. and body parts arms, waist, shoulders, legs etc..

Real Estate & Home Deco : Digital visits and try furniture at finger tips - refurnish your house from your couch Scan Rooms - Houses Virtual Tours, Measurements and Refurnishing

Scanm B.V.: Real Estate: Low-Cost Solution! Scan Rooms - Houses Equipment 3D Scanning for Everyone Lenovo + Rotator

Deep Learning in Computer Vision Object classification, detection and segmentation Optical flow Color constancy Intrinsic image decomposition 3D and slam Human behavior analysis Etc