Introduction Augmented Reality is a view of a real‐world environment that has been augmented by computer input. Motivation: Seemed like an interesting field Becoming more common Enjoy playing chess Project Goals Augment images of a game of chess Decided to draw possible moves for chess pieces Would like to consider rules of chess as well Originally considered the idea of real‐time implementation Previous Work Two main resources Gambit: A Robust Chess‐Playing Robotic System (3D) Locate the board Piece Recognition Square Detector (HOG features) Piece Detector Color Detector Piece Classifier (SIFT) Game Playing Previous Work Shogi Project (2D) Pose Estimation Algorithm Used four colored markers on corners Template Matching to Identify Pieces NCC with sample images Peaks assumed to be piece locations Augment Arrows on Pieces Implementation Camera Calibration Obtain Intrinsic Parameters Pose Estimation Used same method as Shogi Project Piece Recognition SIFT Template Matching Camera Calibration Matlab’s camera calibration toolbox 10 pictures of chess board Originally used 5 Focal Length: fc = [ 2736.25403 2761.31401 ] ± [ 19.46801 19.78666 ] Pixel error: err = [ 1.75785 1.70676 ] Camera Calibration Extracted corners 200 Yc (in camera frame) 400 600 800 1000 1200 1400 1600 dY O dX 1800 500 1000 1500 Xc (in camera frame) 2000 2500 Pose Estimation Based on finding four colored markers on the corners of the chess board Used pose estimation algorithm from class once location of four corners was known Theoretically easy method of automatic pose detection Proved to be annoying Values for colors change largely from small changes in lighting or angle Poor detection while keeping thresholds constant, only 4 out of 8 images correctly determined Pose Estimation (Bad Example) Pose Estimation (Good Example) Piece Recognition Started with an attempt to use SIFT Very little progress despite many attempts Changed code to match from multiple training images Attempted angled views and overhead views Attempted to crop images for easier matching Decided to try other methods after multiple failures Template matching (NCC) Piece Recognition 4 4 1 8 3 3 5 2 6 6 7 8 2 7 1 5 Piece Recognition 2 1 5 3 6 3 1 4 6 7 7 4 5 2 Piece Recognition 5 5 4 3 2 4 1 2 3 1 Piece Recognition Second method: Template matching using normalized cross‐correlation Currently in the process of implementing First attempt caused zero potential interest points to show up Probably need to look at thresholds Final Result Biggest obstacle is 3D object recognition Overhead view oddly difficult for chess pieces Pose estimation based on colored markers proved to be a hassle Areas to Improve Work on color algorithm or try corner/Hough methods 3D Piece Recognition Implement AR Functionality Questions?