Cloud intelligence for mining criminal justice data

Criminal justice organisations worldwide are capturing ever increasing amounts of video and audio footage. Law enforcement agencies are collecting video from body-worn cameras, in-vehicle cameras, and surveillance cameras placed throughout their communities. Detectives are videotaping interviews with suspects. Emergency officials are recording 999 calls. And judicial agencies are making audio recordings of court proceedings.

With hundreds of thousands of hours of video and audio being generated annually and retention policies that often exceed years, agencies are looking to the cloud to cost-effectively store, manage, and protect all of this content.

With the low storage costs and regulatory compliance commitments for UK Government OFFICIAL data; including meeting the 14  cloud security principles, with the Microsoft Azure offering on the G-Cloud framework, UK law enforcement organisations are turning to Microsoft to store their video in Azure.

Hidden in all of this camera and audio footage is a wealth of evidentiary information that could be potentially unlocked for use by detectives investigating crimes for proactive terrorism response and crime fighting. Data points like faces, spoken words, physical objects, license plates, human emotion, and sounds like gunshots or explosions, are all key pieces of evidence that could be used by law enforcement.

Until the inception of the cloud however, much of this evidence was overlooked due to the sheer volume of the content and the number of hours required to manually extract and tag this evidence.

With the new cloud-based cognitive services capabilities in Microsoft Azure, police and the courts can now leverage machine learning, artificial intelligence, and vast amounts of compute power to automatically analyse the video they collect. These API based services can identify and tag faces, redact content to protect citizen privacy, recognise emotion, and identify and tag objects in a video . Using these features of Azure, for example, law enforcement agencies can automatically be alerted to packages or bags left on a street corner for a long period of time as well as unusual noises such as gunshots, explosions, and screeching tires. They can also correlate witness testimony with video footage—for example, the services could automatically search surveillance footage for a “a suspect seen wearing a red jacket.”

Likewise, courts administrators or prosecutors can tap into the audio-indexing features of Azure to automatically turn spoken words into written transcripts. They can also search audio and video recordings for key words to quickly obtain the information they need. For example, lawyers can search for a specific docket number across video footage recorded on multiple days, to obtain the exact audio and video clips they need. Similarly, detectives can search video and audio content for names, addresses, and other keywords, quickly obtaining information about a specific person or location, and identifying potential links between cases.

These are transformative capabilities that can help criminal justice organisations operate more efficiently and improve their ability to fight crime. Take the state of New Jersey, for example. Voters there recently passed bail reform legislation that requires defendants who are accused of an indictable offence to receive a bail hearing within 48 hours of their arrest. To help the state meet these new requirements, Microsoft has been working with several counties to test our audio indexing features.

One factor that makes it difficult to hold prompt bail hearings in New Jersey is that it can take from three to six hours to transcribe one hour of audio or video content. And with 999 calls, video captured at the scene, and interviews of multiple suspects, that can add up to a lot of transcription time. The audio-indexing features of Microsoft Azure automatically transform spoken content into written transcripts in a matter of minutes, greatly reducing costs and the work of stenographers who simply need to review these transcriptions for accuracy.

Cognitive services capabilities are poised to transform how criminal justice organisations worldwide investigate crimes, prevent terrorism, improve public safety, and document criminal justice proceedings. With the petabytes of data now being stored in the cloud, the next logical step is for agencies is to understand how they can use this to their advantage. We are making deep investments in our cloud cognitive services engines and are working alongside agencies to deploy a smarter, proactive model for criminal justice data mining.

To learn more about how Microsoft Azure can extract value from public safety data, please see our Microsoft Cognitive Services webpage.

To find out more about mining data, download our eBook ‘Your guide to unlocking organisational growth with data’