Since 2008 Google street view has blurred faces for privacy.

Slides:



Advertisements
Similar presentations
The following slides show different approaches to completing page 11.
Advertisements

Strong red lines surround a human figure, echoed in the repeated human figures below. The warm (red lines) and cool (blue background and dancers) colors.
Wen-Hung Liao Department of Computer Science National Chengchi University November 27, 2008 Estimation of Skin Color Range Using Achromatic Features.
Simple Face Detection system Ali Arab Sharif university of tech. Fall 2012.
COLORCOLOR A SET OF CODES GENERATED BY THE BRAİN How do you quantify? How do you use?
F ACE TRACKING EE 7700 Name: Jing Chen Shaoming Chen.
© red ©
1 Color Segmentation: Color Spaces and Illumination Mohan Sridharan University of Birmingham
CHEN Guowang FANG Wei HUANG Baihan
Classification Dr Eamonn Keogh Computer Science & Engineering Department University of California - Riverside Riverside,CA Who.
Color: Readings: Ch 6: color spaces color histograms color segmentation.
Smart Traveller with Visual Translator. What is Smart Traveller? Mobile Device which is convenience for a traveller to carry Mobile Device which is convenience.
Rugby players, Ballet dancers, and the Nearest Neighbour Classifier COMP24111 lecture 2.
Color Space for Skin Detection – A Review
CS 101 – Sept. 16 Finish color representation –RGB √ –CMY –HSB –Indexed color Chapter 4 – how computers think –Begin with basic building blocks.
Facial Feature Extraction Yuri Vanzine C490/B657 Computer Vision.
Remote Sensing and Image Processing: 2 Dr. Hassan J. Eghbali.
Algirdas Beinaravičius Gediminas Mazrimas Salman Mosslem.
DIEGO AGUIRRE COMPUTER VISION INTRODUCTION 1. QUESTION What is Computer Vision? 2.
Taking Better Confocal Images Kim Peifley 2/09/15.
44212: Web-site Development Optimising Images for Web-pages Ian Perry Room:C49 Extension:7287
By: Jared Meyer A Neural Network Approach to Classifying Cartoons Based on Color ECE 539 Final Project.
W2D1. HSV colour model Source: Primary colours Red Green Blue Secondary Cyan Magenta.
Histograms and Color Balancing Computational Photography Derek Hoiem, University of Illinois 09/10/15 “Empire of Light”, Magritte.
Digital Image Processing In The Name Of God Digital Image Processing Lecture6: Color Image Processing M. Ghelich Oghli By: M. Ghelich Oghli
Clustering Prof. Ramin Zabih
Color Color Color Tsung-Yi Wu.
` Tracking the Eyes using a Webcam Presented by: Kwesi Ackon Kwesi Ackon Supervisor: Mr. J. Connan.
Computer Vision Spring ,-685 Instructor: S. Narasimhan WH 5409 T-R 10:30am – 11:50am Lecture #23.
HW6: Shot Detection B 劉品賢. Method Histogram differential Part 1 Using RGB color space Part 2 Using YCbCr color space but only analysis Y color.
Sensing Colors. B G Color Digital Image R Red sensor Green sensor Blue sensor.
EE368 Final Project Spring 2003
Elements & Principles of Design
CS262: Computer Vision Lect 06: Face Detection
Sensors Color Identification Mike Zook 13-Sep-2016.
Histograms CSE 6363 – Machine Learning Vassilis Athitsos
Color Image Processing
Color Image Processing
Binary Representation in Audio and Images
Processing the image with Python
COLOR space Mohiuddin Ahmad.
Images In Matlab.
Tutorial 4 Topic: CSS 3.0 Li Xu
Computer Graphics Different Images File.
AUTOMATIC IMAGE ORIENTATION DETECTION
Session 7: Face Detection (cont.)
*Currently, self driving cars do a bit of both.
Color Image Processing
Color: Readings: Ch 6: color spaces color histograms
Investigating Color In the Classroom
Color Values All colors in computer images are a combination of red, green and blue Each component is encoded as a number means the color is.
Lecture 25: Introduction to Recognition
Aberdeen Grammar School
Sensors Color Identification Mike Zook 13-Sep-2016.
Color: Readings: Ch 6: color spaces color histograms
What do these words mean to you?
Color Image Processing
Name: _______________________________
Segmentation of Images By Color
The RGB LED.
Basics of Classification
Light Light has wave-like properties
Estimation of Skin Color Range Using Achromatic Features
Color Image Processing
What Color is it?.
Face Detection in Color Images
Tracking the Eyes using a Webcam
Building Complex Behaviors: Actions and States
Help Gina and Tina color the numbers
Let’s Learn the Basic Colors
Presentation transcript:

Since 2008 Google street view has blurred faces for privacy.

The system mostly works well, but does have both false positive and false negatives

Let us see if we could solve this problem, with what we have learned in this class. First, a small insight. Let us solve the skin detection problem, instead of the face detection problem. What features can we use? Every color pixel in an image, is represented as a combination of three colors, typically Red Green Blue, but there are other possibilities, HSV (hue, saturation, value), HSL (hue, saturation, lightness/luminance) etc. R B G

R B G

However, this rule is not perfect (next slide) B G R = If we project a skin pixel into 3D space, then you could use the following, human coded rule to classify skin pixels: If (R > 95) AND (G > 40) AND (B > 20) AND (max{R, G, B} − min{R, G, B} > 15) AND (|R − G| > 15) AND (R > G) AND (R > B) Then IsSkin However, this rule is not perfect (next slide) G = B =

The rule works pretty good here, but it has a lot of false positives, the plate on the wall, the stripy sweater, parts of the couch on the bottom left etc.

Let us try to do better, with machine learning. First we need to get training data to build our classifier. Recall that for machine learning problems, it can sometimes be useful to transform features. Instead of height and weight, use weight/ height2 Maybe there is a transformation of the color space that makes this problem easer? We could ask an expert, or just compare all color spaces using leave-one-out classification.

 In YCbCr, the Y is the brightness (luma), Cb is blue minus luma (B-Y) and Cr is red minus luma (R-Y). Y = 50 100 150 200 250 Cr Cb = Cb Cr =

Cr > 124? Cr < 174? Not Skin Cb < 90? Not Skin Cb > 140? 50 100 150 200 250 Cr Cr < 174? Not Skin Cb Cb < 90? Not Skin Cb > 140? Not Skin Skin Not Skin

50 100 150 200 250 Cr Cb RGB-H-CbCr Skin Colour Model for Human Face Detection.Nusirwan Anwar Abdul Rahim, Kit Chong Wei, John See.MMU International Symposium on Information & Communications Technologies (M2USIC 2006),

This is a hand coded piecewise linear model Cr ≤ 1.5862 × Cb + 20 Cr ≥ 0.3448 × Cb + 76.2069 Cr ≥ -4.5652 × Cb + 234.5652 Cr ≤ -1.15 × Cb + 301.75 Cr ≤ -2.2857 × Cb + 432.85 50 100 150 200 250 Cr Cb RGB-H-CbCr Skin Colour Model for Human Face Detection.Nusirwan Anwar Abdul Rahim, Kit Chong Wei, John See.MMU International Symposium on Information & Communications Technologies (M2USIC 2006),

A simple nearest neighbor algorithm in the YCbCr space works very well. How could we make it better?

Extensions We can use basic classifiers as a building blocks for more complex questions For example, if we wanted to find examples of FedEx trucks on streetview, we could build a classifier for each color in the logo (Purple, Orange and White). If a small area of the image had all three colors, we could classify truck seen. Why not just “read” the text? Consider the below…

Less we get too complacent I How well could you do on this problems? Google “Karen Zack” for more examples

Less we get too complacent II How well could you do on this problems?