Local authorities needing to review hours and hours of video footage could find help with rapid video review technology. One company making such software is Kinesense, an Irish tech company that has been helping police forces around the world, including the UK, to analyse video for more than ten years.
At its core, Kinesense LE analyses recorded video, applying algorithms to detect movement, colour and objects, with optional facial recognition tools also available. On top of this, it adds filter options, reporting tools and export and image enhancement functions to create a comprehensive package for the video investigator.
I caught up with Vernon Pratt at Kinesense who explained that while there are other rapid video review systems available, what has typically prevented local authorities from investing in them is the cost, particularly the need to purchase very high specification computer hardware to run them. He says Kinesense runs extremely well on good-spec, off-the-shelf PC hardware, putting it within the budget of police forces and local authorities.
NEW Watch our Snapshot Webinar: Rapid video review in the CCTV control room, with Kinesense
And as requested during the webinar, here's the company contact details:
Website: www.kinesensevca.com
The video review challenge
Local authority CCTV control rooms generate thousands of hours of video footage each day. When there is an incident to investigate, it means manually reviewing hours of video to find a few evidential clips. And it doesn’t have to be a major crime, either – it could be an insurance claim, anti-social behaviour, dog fouling, fly-tipping or other low-level crime.
Councils facing increased demands from their residents can encounter challenges in the preparation of evidential quality evidence from CCTV video, including searching the video for events, extracting high-quality video and still images and preparing intelligence, court and disclosure reports.
“How many of these incidents aren’t investigated because there simply isn’t time or resources to review the video?” Vernon asks. And with the increased use of rapid deployment cameras to identify fly tippers and other types of anti-social behaviour, how can councils streamline the capture of evidential video from these devices?
Kinesense could change the way you are able to use the hours of video footage in your systems, he says.
CCTV managers and consultants that I have spoken to say that rapid review video would be particularly useful in the investigation of environmental crime such as fly tipping, offences which are typically left to the environmental enforcement team to pursue as civil matters.
In addition, police are frequent users of local authority control rooms. Video review suites are often provided for them, but what if these suites were equipped with rapid review technology? Rapid review software could reduce the amount of time that police officers need to spend in control rooms, or even allow police officers to perform some of the work remotely (with a little assistance from the control room staff).
And, as one CCTV manager suggested to me, given the move by police to promote the remote ingestion of council-generated CCTV content (using online or cloud-based Digital Evidence Management Systems (DEMS)), CCTV operators can speed up the search for any offences that have been captured prior to uploading for police review.
How Kinesense works
Kinesense is a full-featured video investigation platform which enables investigators to use algorithms to review video and then generate reports without having to switch between multiple applications – in other words, no need to cut and paste still images into an Excel document.
It begins with an ability to ingest video from multiple sources. In fact, Kinesense claim that their solution, with 900 built-in virtualised players, has the ability to take in video from the majority of the 4,000 different video formats available, including dashcams, mobile phones, domestic grade CCTV, covert cameras and other sources. While this has obvious advantages for police, councils are also increasingly deploying new types of camera technology along with VMS such as Milestone and Genetec, which can also feed directly into Kinesense, a feature that could be relevant in many areas.
The core feature of Kinesense LE is the automated event detection and analysis, to which you can apply customisable filters. More on this below.
Kinesense can also integrate audio files from emergency, customer service and out of hours helpline numbers and has a map-building feature which allows an investigator to link evidential clips to mapped locations and display associated notes and tags.
After reviewing, tagging and adding notes to clips, the clips can be assembled in a storyboard and exported to reports. This includes an automatically generated disclosure report which details the actions that have been applied to each clip including filters and enhancements, creating an audit trail that is compliant with ISO 17025.
Kinesense filters
As mentioned, the core features of Kinesense LE are the automated event detection and customisable search filters which can enhance the effectiveness of search while cutting down the amount of labour required to do it.
Vernon took me through a demo of the system, explaining that video analysis begins by ingesting video and running algorithms against the images. A key feature is that it does not modify the source video – all data and enhancements are stored in a separate, indexed database, preserving the evidential integrity of the original video.
After ingesting video, Kinesense analyses the footage to identify scene changes. This is displayed in a graphical timeline as orange dots in the blue timeline below the video.
Kinesense filters can identify:
Colour
Direction of travel
People v vehicles
Zones of interest – eg, a doorway or section of pavement
Optional face recognition is available
Once video has been ingested and analysed, filters can be adjusted without the need to re-ingest or re-analyse the video so if you need to fine tune a filter to cut out false negatives or positives, it won’t take hours to reanalyse the video, Vernon says.
Clearly, without using filters, you would get a lot of notifications, especially if your video included a busy road, so one of the most useful filters to start with is zone masking. Zone masking allows you to exclude any movement outside your area of interest.
In the example in this image, the investigator has masked in an area at the bottom right of the frame, between the car, the house and the wheelie bin (outlined in orange). The immediate effect of this is to reduce the number of orange activity indicators in the blue timeline.
With this, the investigator can quickly find all video clips showing someone entering or leaving the house. Click a button (Play Events Only), and the player will automatically play only the highlighted clips (with a brief pre-roll and post-roll for context).
By playing only those clips that contain events in the zone of interest, you can cut hours off video reviews. Built-in reporting tools allow you to add tags and notes to clips, creating the data for your evidence report as you go.
Evidential video tools
Having reviewed the clips, adding tags and notes, you can export reports, including a storyboard, tags report and disclosure statement showing what was done to the video.
The automated tools eliminate the need to copy and paste images and data into Excel spreadsheets and is ISO 17025 compliant.
The technology has already been tested in court, with Judge Richard Mansell QC and defence counsel in one case accepting that viewing video footage with dedicated software constituted independent review.
In addition, the software contains video enhancement and annotation tools which allow you to:
Redact moving people from clips easily with tracking masks
Add text and arrows which track movement
Improve image quality with deblur, skew and contrast tools
All actions applied to an image, from enhancement to annotation, are logged in the disclosure report.
Mapping tools can show:
Camera location and view
Associated video evidence
Tags including vehicles and individuals
The benefits
Vernon says local authorities can benefit from many of the same features of Kinesense that police forces are currently enjoying:
Find evidence up to 20 times faster
Reduce reporting time by more than 40%
Motivate staff with faster, less onerous tools
In addition, the image redaction tools make generating subject access requests faster and more efficient, he says, and creating evidence packages for investigations becomes quicker and less error prone.
For the police, it enables video to be analysed at a more local level, significantly reducing the workload on digital forensics units (DFU). In Northumbria, for instance, it has been given to 13 local police stations as well as the DFU, reducing their workload from 7,000 cases per year to just 600. This has cut the DFU backlog from months to weeks, and it has also significantly reduced storage costs for the force by saving only evidential clips.
For a CCTV control room, Kinesense LE can be set up in a review suite as a tool for the council or the police. It can be linked directly to your recording systems, reducing the need to copy extensive amounts of video footage to a workstation.
Kinesense significantly speeds up finding key evidence, Vernon says. It eliminates the boring and onerous tasks associated with video review, allowing skilled operators to concentrate on the key tasks of reviewing and logging clips with potential evidential value. It is an integrated tool, eliminating the need to copy information between different software packages and significantly reducing the scope for human error, he adds.
Kinesense also provides a digital evidence management system (Kinesense DEM). If your local authority or regional police force has this system, it would enable you to upload evidence packages directly, streamlining the management of evidence between you, the police and the Crown Prosecution Service – saving time and the use of temporary storage media.
Learn more about Kinesense by attending the CCTV User Group Snapshot Webinar “Rapid Video Review with Kinesense”. In this live webinar at 2.30pm on 10 December, I’ll be speaking to Vernon Pratt about Kinesense rapid video review technology. There will be a short presentation followed by a Q&A session during which those attending the live event can ask any questions. To join us, register here.
Written by: Tom Reeve, Chief Communications Officer, CCTV User Group
コメント