Crime Forecasting Using Historical Crime Location Using CNN-Based Images Classification Mechanism

Abstract

The present research is focused on crime forecasting and CNN has been used for image classification in order to categorize crime events. However, there are different classification mechanisms used in conventional research work. But CNN is playing a significant role in identification and prediction of crime. The major issues during CNN based classification are time consumption and accuracy. However, proposed research has resolved issue of time consumption by reducing image size by applying the RGB2GRAY model, and images are resized before training operation. Simulation results conclude that the proposed work provides a scalable and reliable approach for crime forecasting.