Crime Detection and Evidence Extraction Using Machine Learning on Cloud

Digital video plays a vital role in evidence identification, analysis, presentation, and report recently. With the convenience use of smartphones and the increasing popularity of surveillance camera, visual data are hugely being used in digital crime investigation. Extracting evidence becomes a time-consuming process when searched manually. To ensure reduction in possible human errors and for better accuracy, machine learning model is used which speeds up the investigation by extracting crime scene objects and masked faces from the crime scene footages. The main approach of this paper is to develop such machine learning model to assist forensic investigation for better evidence extraction with higher accuracy. In order to further reduce the processing time of the object extraction model, parallel processing is introduced which helps the model to perform faster in larger data sets. The evidence extracted from the video are stored in the cloud storage repository. Storing of these extracted evidences in the cloud helps to compare the extracted evidences with the past crimes and these extracted crime scene objects can be used to train the model further in order to improve the detection rate.