I need suggestions for my project and would be glad if you would give me a hand.

I have a dataset of frames obtained from the old-school game DOOM. Each frame in the dataset has the following columns:

time, frame buffer size in bytes, game stage (0: Exploration, 1: Combat)

Once collecting consecutive frames to calculate instant bitrate (kbps) for each time instance, I apply Exponential Moving Average on bitrate to smooth and here below the data I end up with:

Bitrate Data

I need to train a model by using this data and in the end, I should be able to classify game stages correctly. Currently, I have two approaches:

  1. Calculating the average bitrate of each stage, and examining the % change from the former stage to the latter one (avg_latter - avg_former / avg_former * 100) in different baseline bitrate ranges: (W=9, 18, 35 below indicates Exponential Moving Average window lengths)

Result for transitions from Exploration -> Combat

Result for transitions from Combat -> Exploration

  1. Calculating the bitrate variance in each stage change moment by looking 'N' frames back and forward: Bitrate Variance Calculation

Results of transitions from Exploration -> Combat and viceversa

According to the results above, I wonder which methodology would make more sense for game stage classification. Any idea is appreciated. Thanks!


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.