100 likes | 239 Views
A Neural Network Approach to Classifying Cartoons Based on Color. ECE 539 Final Project. By: Jared Meyer. Project Plans. Collect Data Choose 20 different animated series Choose 3-5 episodes at random Varying seasons if applicable Covert each episode to series of images
E N D
A Neural Network Approach to Classifying Cartoons Based on Color ECE 539 Final Project By: Jared Meyer
Project Plans • Collect Data • Choose 20 different animated series • Choose 3-5 episodes at random • Varying seasons if applicable • Covert each episode to series of images • Calculate data for each image • Write program in C# • Build Artificial Neural Network with bp.m • Find structure/data combination that maximize classification rate
Why? • Big fan of cartoons/animated series • Interested in how images are represented in computers • Would be neat to see color patterns in shows • Existing Results • Weather classification based on color (Moosmann, 2008) • Linear kernel Support Vector Machine • 3 classes • Clear • Light Rain • Heavy Rain • Average classification rate: 89%
Project Steps • Ripped 3-5 episodes of following shows: Avatar: The Last Airbender The Real Ghost Busters Batman Reboot Courage the Cowardly Dog Samurai Jack Cowboy Bebop The Simpsons Ed, Edd, n’ Eddy South Park Family Guy Spiderman FuturamaSpongebobSquarepants Invader Zim Superman Outlaw Star Teenage Mutant Ninja Turtles Powerpuff Girls Teen Titans These form the 20 outputs for ANN
Project Steps • Converted episodes to images • X Video Converter • One BMP image per 200 frames • Remove first frame • Usually pure black • Remove all frames including end credits • Would add bias
Data Calculations • Wrote program in C# to calculate 14 Features per image • Brightness, Contrast, Saturation, RGB ratio • ‘Lininess’ • Pixels with large brightness difference
Data Calculations • ‘Important Areas’ • Pixels brighter than average brightness • Counted Red, Orange, Yellow, Green, Blue, Violet, Grey Pixels in ‘Important Areas’ • Finally, used bp.m program to build ANN using back-propagation algorithm
Results • Data varied greatly, even in same episode • ~5% classification rates • Averaged 10 random frames together
Results • Contrast, Color counts still varied too much • Removed them; didn’t show much pattern
Results • Much better classification rates with new data Final Results: 5 Features: Brightness, Saturation, RGB ANN Structure: 2 Hidden layers, 9 neurons Classification Rates: 57.14% on Training 47.50% on Testing Pretty good, considering we had 20 classes.