CV Dazzle

CV Dazzle is camouflage from computer vision (CV). It is a form of expressive interference that combines makeup and hair styling (or other modifications) with face-detection thwarting designs.

The name is derived from a type of camouflage used during WWI, called Dazzle, which was used to break apart the gestalt-image of warships, making it hard to discern their directionality, size, and orientation. Likewise, the goal of CV Dazzle is to break apart the gestalt of a face or object, and make it undetectable to computer vision algorithms, in particular face detection.

For more information visit CVDazzle.com

Published May 2010
CV Dazzle Look #2
CV Dazzle Look #2

From the DIS Magazine story “How to Hide from Machines

OpenCV vs CV Dazzle / 2010
OpenCV vs CV Dazzle / 2010

Test results of Dis Magazine’s style against OpenCV

CV Dazzle Look #1
CV Dazzle Look #1

First test pattern against OpenCV

OpenCV Vs Cv Dazzle / 2010
OpenCV Vs Cv Dazzle / 2010
CV Dazzle vs OpenCV / 2010

Test results of first camouflage design pattern against OpenCV. Hair styling by Pia Vivas

CV Dazzle Look #3
CV Dazzle Look #3

From the DIS Magazine story “How to Hide from Machines

CV Dazzle Look #4
CV Dazzle Look #4

From the DIS Magazine story “How to Hide from Machines

Dis Magazine + CV Dazzle. Photo by Marco Roso.
Dis Magazine + CV Dazzle. Photo by Marco Roso.

From the DIS Magazine story “How to Hide from Machines

Behind the Scenes at the DIS + CV Dazzle Collaboration / 2010
Behind the Scenes at the DIS + CV Dazzle Collaboration / 2010

From the DIS Magazine story “How to Hide from Machines

 
 

OpenCV Face Detection: Visualized from Adam Harvey on Vimeo.

This video visualizes the detection process of OpenCV’s face detector. The algorithm uses the Viola Jones method of calculating the integral image and then performing some calculations on all the areas defined by the black and white rectangles to analyze the differences between the dark and light regions of a face. The sub-window (in red) is scanned across the image at various scales to detect if there is a potential face within the window. If not, it continues scanning. If it passes all stages in the cascade file, it is marked with a red rectangle. But this does not yet confirm a face. In the post-processing stage all the potential faces are checked for overlaps. Typically, 2 or 3 overlapping rectangles are required to confirm a face. Loner rectangles are rejected as false-positives.

Test Patterns for CV Dazzle
Test Patterns for CV Dazzle
  • Images with a red square tested positive, a face was found
  • Images without a red square tested negative, no face was found
  • Images under the section “TEST PATTERNS” are made according to results of the Haar deconstruction
  • Images under “RANDOM PATTERNS” are random doodles made without the anti-face detection patterns in mind
  • Images underneath the “NO PATTERNS” heading are left untouched to show that the face detection works well on simple line drawings
  • Line drawings are from Figure Drawing for Fashion Design
Visualization of Stage 3 of Haarcascade
Visualization of Stage 3 of Haarcascade

Visualization of a Haar Cascade profile for OpenCV face detection. The face detection algorihtm looks for the difference between the light and dark regions and determines whether it is within the threshold of a face pattern. Each cascade contains over 20 stages. Shown here is stage 3. Each stage becomes more discerning so that non-faces are quickly rejected in the beginning of the image scan.