Download presentation
Presentation is loading. Please wait.
Published byBranden Sullivan Modified over 9 years ago
1
Android QR-Code Detection Cerman Martin, 0625040 mcerman@prip.tuwien.ac.at
2
Content Topic & Challenges (original proposal) IDE‘s, Languages & Libraries Chosen Approach Outlier Filtering Tuneable Parameters TODO (original proposal / what was achieved) Topic & Challenges (what was achieved) Live Demonstration Android QR-Code Detection, Cerman Martin, 06250402
3
Topic & Challenges (original proposal) Detect and read QR-Codes on an Android Phone As defined by the ISO/IEC Standart „Set of black squares on a white background“ Various sizes (Version 1-40) Android QR-Code Detection, Cerman Martin, 06250403 Topic Detection under various lighting conditions Perspective distortion Detection of QR-Code size Real-Time Challenges / Goals
4
IDE‘s, Languages & Libraries Android QR-Code Detection, Cerman Martin, 06250404 MATLAB Eclipse ADT (Android Development Toolkit) CDT (C++ Development Toolkit) NDK (Android Native Development Kit) Visual Studio 2012 OpenCV Library
5
Chosen Approach Android QR-Code Detection, Cerman Martin, 06250405 „Fast Radial Symmetry for Detecting Points of Interest“, [1] Algorithm Transform to grayscale and reduce image size Determine gradient in x and y direction Compute orientation and magnitude image Compute symmetry image at certain radii Smooth using a Gauss kernel Non-maximum suppression
6
Chosen Approach (transform to grayscale) Android QR-Code Detection, Cerman Martin, 06250406
7
Chosen Approach (x derivative) Android QR-Code Detection, Cerman Martin, 06250407
8
Chosen Approach (y derivative) Android QR-Code Detection, Cerman Martin, 06250408
9
Chosen Approach (orientation image) Android QR-Code Detection, Cerman Martin, 06250409
10
Chosen Approach (magnitude image) Android QR-Code Detection, Cerman Martin, 062504010
11
Chosen Approach (symmetry image) Android QR-Code Detection, Cerman Martin, 062504011
12
Chosen Approach (thresholded symmetry image) Android QR-Code Detection, Cerman Martin, 062504012
13
Outlier Filtering Android QR-Code Detection, Cerman Martin, 062504013 Maximal number of outliers is around 20 Algorithm Build a complete graph Compute for each vertex distances to all other vertices Compute angle for each vertex trio Filter outliers by setting restrictions on maximal angle deviation and distance difference for each vertex trio Turned out to work very well
14
Tuneable Parameters Android QR-Code Detection, Cerman Martin, 062504014 Feature Detection Image reduction factor Search radius Minimal number of positively affecting pixels Symmetry strength Outlier Filtering Maximal edge length difference of each vertex trio Maximal angle deviation of each vertex trio Frame Management Split up searching (different scales) and filtering among frames
15
TODO (original proposal) Android QR-Code Detection, Cerman Martin, 062504015 Determine parameters Image size Radii Filter outliers Based on neighborhood Find suitable feature descriptor QR-Code reading
16
TODO (what was achieved) Android QR-Code Detection, Cerman Martin, 062504016 Determine parameters Image size Radii Filter outliers Based on neighborhood Find suitable feature descriptor QR-Code reading
17
Topic & Challenges (what was achieved) Detect and read QR-Codes on an Android Phone As defined by the ISO/IEC Standart „Set of black squares on a white background“ Various sizes (Version 1-40) Android QR-Code Detection, Cerman Martin, 062504017 Topic Detection under various lighting conditions Perspective distortion Detection of QR-Code size Real-Time Challenges / Goals /
18
Live Demonstration 18your title
19
References Android QR-Code Detection, Cerman Martin, 062504019 [1] Fast Radial Symmetry for Detecting Points of Interest, G. Loy and A. Zelinsky, IEEE Transactions on Pattern Analysis and Machine Intelligence, August 2003
20
Thank you
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.