Context Wondering if a model can detect visual differences through gaming graphics? If it is possible, does the model have real world applications? Check out these images from Virtual Battle Station 2 and see if you can build the model and identify any further trends in this visual dataset. Content The dataset consists of before and after scenes with one major difference (new building, road, bush or car) and lots of smaller minor differences (lighting, weather, ...). The goal is to automatically detect the major differences among the noisy background of other smaller differences. Acknowledgements The original dataset page is down but the full dataset can still be downloaded here: https://computervisiononline.com/dataset/1105138664 Credit: Nicolas Bourdis, Denis Marraud, Hichem Sahbi Inspiration How well do simulated video game images apply to real world problems?