Authors:
Matthieu Chan Chee
1
;
Vinay Pandit
2
and
Max Kiehn
2
Affiliations:
1
University of Toronto, Ontario, Canada
;
2
AMD, Inc., Thornhill, Ontario, Canada
Keyword(s):
Video Game Display Corruption, Image Corruption Detection, Deep Convolutional Neural Network, EfficientNet, Structural Similarity Index Measure, Grad-CAM.
Abstract:
Early detection of video game display corruption is essential to maintain the highest quality standards and to reduce the time to market of new GPUs targeted for the gaming industry. This paper presents a Deep Learning approach to automate gameplay corruption detection, which otherwise requires labor-intensive manual inspection. Unlike prior efforts which are reliant on synthetically generated corrupted images, we collected real-world examples of corrupted images from over 50 game titles. We trained an EfficientNet to classify input game frames as corrupted or golden using a two-stage training strategy and extensive hyperparameter search. Our method was able to accurately detect a variety of geometric, texture, and color corruptions with a precision of 0.989 and recall of 0.888.