Video Encoding Energy and Greenhouse Gas Emission Prediction

Abstract

Techniques relating to video encoding energy and greenhouse gas emission prediction are disclosed. A system for video encoding energy and greenhouse gas emission prediction includes a video analyzer, an energy predictor module, a carbon data source, and a carbon emissions calculator. The system may be configured to carry out a method that includes extracting complexity features from a video segment, predicting energy consumption for video encoding the video segment on a cloud instance using a machine learning (ML) model, receiving carbon intensity data, cloud instance type data, and region (i.e., country or set of countries) data, and calculating greenhouse gas (e.g., CO2) emissions based on predicted energy consumption and carbon intensity data. Both fossil fuel and renewable energy sources may be accounted for, along with power imports and exports using a peer-reviewed flow-tracing methodology. The ML model may use one or both of ensemble-based and linear regression methods.

CPC Classifications