This document summarizes a research paper that proposes an audio-based approach for retrieving action events from cricket videos. The approach detects action events by measuring abnormal increases in audio energy levels during batsman strokes and crowd cheers. The researchers extract audio features like MFCC coefficients and calculate audio energy values from cricket video soundtracks. Peaks in energy levels are identified using adaptive thresholding to detect strokes and cheers. Video frames around detected audio peaks are retrieved as highlights of the action event. The method was tested on a dataset of cricket videos and results showed it can efficiently retrieve events like strokes and crowd reactions.