Paper Title
Introducing A Voice Detection System for Criminal Investigation Unit

Abstract - The similarities of a defendant's personal characteristics to those of the perpetrator of a crime can be used to establish legal guilt. Eyewitness testimony is often used to make such comparisons, and eyewitness testimony implicates dialects and language behavior. Furthermore, since voice is a personal characteristic on which a criminal's identification is based, it could be at issue in a criminal case. The perpetrator's voice may be only clear evidence available for police investigations and convictions in obscenity, bomb hoaxes, ransom claims, captured violations, robberies, muggings, or crimes committed in the dark. The way that most exploration has been done in the field of witness declaration and ID shows that most identifiers include an observer using visual signs. Therefore, recognizing voices also becomes important as well as the eyewitness which will help to identify criminals. As a solution, we are introducing a system that engenders voice patterns utilizing each entity's unique factors like word gaps when verbalizing, dialects, the structure of verbalization, and the ways a sentence is engendering using voice recordings of criminals and stores it in a database.The modern machine learning and deep learning algorithms such as butterworth, high-pass, lowpass have been used by this system to reduces the background noise. The voice pattern of a particular utilizer will be compared with the database which contains all voice patterns and if they ascertain any kindred pattern belongs to that utilizer, an alert will automatically engender and sent to the nearest Police Station. So, it will be very convenient for Criminal Investigation Unit to identify malefactors and their adherents utilizing the database. Keywords - Criminal, Voice Patterns, Machine Learning, Deep Learning, Butterworth, High Pass, Low Pass