AiMS —Meet the New VMS to Reduce Kidnapping with AI Video Analytics
With the gap between the rich and poor is at its highest level in decades, the rising kidnapping rate has drawn the attention of the public and governments worldwide. Concern for the inefficiency of investigation is growing, and the difficulty of quickly finding out the clues of suspects in time has become one the most significant challenges for the authorities.
In recent years, Artificial Intelligence technologies such as human detection and facial recognition have proven themselves a highly effective solution to handle these security issues.
SkyREC, as a video analytics software innovator powered by AI, launches the latest video management software (VMS) — AiMS (Autonomous Intelligent Management System) which brings all the puzzle pieces for video surveillance together to create an ultra-robust and super intelligent solution that meets the requirements of all industries.
In short, SkyREC AiMS is a crime deterrent and is now used to collect data points with precise AI video analytics to help detect abnormal behaviors for proactive post-event investigations with high efficiency.
Here is an example of how AiMS takes investigations to new heights in a kidnapping scenario.
We often see news on TV like this to look for help from the public: the police got the following initial clues either based on the impression of the victims or witnesses, such as a child in a green t-shirt was lured into a white sedan by a man. It might take the police months or years for further progress or even end up in vain.
The story above is typical with specific clues but lacks efficient technology to speed up the investigations or prevent the tragedy to happen again, which deteriorates the crimes to some degree.
To address this issue, SkyREC AiMS has evolved from simply facilitating camera and video management to a more sophisticated and advanced role. Object Focus, Object Search, Appearance Focus, and Behavior Focus of AiMS are the main highlights of being a search engine powered by deep learning AI video analytics, enabling the police or security operators to rapidly build and bookmark video clips tracking specific incidents or target objects, reducing the time spent searching for security video evidence from hours to just minutes, or seconds short.
Object Focus
Object Focus performs logical reasoning of the system to provide real-time detection, analyzing, and tracking meaningful objects to different scenarios. As AiMS learns, thinks, and optimizes itself day by day, the accuracy of video analytics continues strengthening on the face, head, human, and vehicle recognition, which greatly benefits the live monitoring and on-demand modes.
Security operators or security guards just put the highlight conditions to AiMS to automatically recognizes the object which is required attention from the live streams of hundreds or thousands of cameras, and Object Focus pops out the notifications by highlighting the target objects without having the security officers stare non-stop at the cameras until people finally lose focus from dozens of monitors.
In a busy train station, if a child or teenager appears on the platform at an inappropriate timing for example late evening, it is very likely this young person is getting into trouble caused by online abduction via social media or trying to run away from home for reasons. As soon as the security operators get an instant alert by AiMS Object Focus, they can inform the station staff to make an approach for preventing possible accidents.
Object Search
To strengthen the very limited visual intelligence capabilities of surveillance cameras (especially regular IP cameras, which are the most common system investments for security purposes due to budget concerns), either in indoor or outdoor scenes, Object Search rapidly identifies specific objects based on people characters including human face, head, and body or analyzes vehicles.
Simply by one click, searching multiple objects simultaneously in any zone of interest coming up with a powerful searching result will be complete within seconds or even milliseconds. Users then can choose suspicious images to bookmark for further video applications such as Appearance Focus or Object Playback. Object Search represents a new way of video search to restore the truth and navigate investigators to the specific moment of an event in seconds.
In the case of kidnapping on the street, many incidents happen outside the campus entrance during school opening or closing hours; some could even occur near the house while the adults are just temporarily away from their beloved kids. Object Search helps the police to quickly look through hundreds or thousands of security videos near the area and lock down the crime suspects along the way of the kidnapping by witnesses’ impressions or photos they took of the suspect or suspicious car.
Appearance Focus
Once the initial lockdown of the target object is complete with Object Search, an innovation searching function, Appearance Focus powered by Deep Learning, provides an ultra-robust search outcome to fulfill its destiny for boosting the investigation efficiency.
SkyREC Appearance Focus achieves complex tasks on video search by enabling user-friendly but detailed visual instructions and feedback. First of all, Appearance Focus enables users to filter hours of security video footage within milliseconds to quickly pinpoint the target object.
In addition, users have great flexibility to customize their search condition combinations by setting critical physical descriptions, including clothing type, clothing color, accessories, face mask, gender, age, moving direction, posture, behavior, vehicle type, or vehicle color categorization. Besides, users are free to upload the picture of the target for an even target-driven search result.
With a more sophisticated search engine design and training materials, Appearance Focus empowers apparel recognition and vehicle recognition models which are standard functions of an AI Video Management Software to be more innovative and agile.
Behavior Focus
To confirm if a series of actions of kidnapping is simple the runaway child, executed by an individual criminal, or planned by a transnational organized crime, it is crucial to track the continuous behavior of the suspect(s). On those post-event investigations requiring more details, SkyREC Behavior Focus outperforms other video management software by bringing superior AI analytics on video recording and playback.
In order to look wider and deeper into the incidents that occur in the same area on different timelines, Behavior Focus not only synchronizes video playback for multiple objects by uncovering the beginning and ending time of each behavior on the time bar but also simultaneously zooms in the details of each specific target without wasting time on opening extra windows, so that the users can quickly figure out how the incident happened and what, when and where to track next.
In the case mentioned earlier, by using SkyREC AiMS, just with a few steps, a missing child who was taken away by the suspects can be quickly and precisely found leveraged by the power of AI Video Analytics.
Firstly, the investigators of police use Object Search or Appearance Focus in the crime scene to find the traces of the missing child at a specific moment by face/human searching, or the apparel search conditions that are reported by the victim’s family or witnesses.
After quickly filtering complex scenes of massive recorded videos, the initial result shows the related images of one suspect and the crime car aside from the missing child. The police then can use clues as to the search conditions under the Appearance Focus model to speed up finding out other associates by searching apparel and vehicle conditions along the way of kidnapping.
Moreover, to expedite reviewing significant videos, the investigators can utilize Behavior Focus that assists them to search persons and vehicles at the same time with one click. Thus, the result is shown in a window that offers the respective time within different time bars. Additionally, it benefits the police to collect more evidence and unveil the details of the kidnapping process to prevent the next tragedy.
During this complicated investigative process, bookmarking the critical images allows the investigators to star important images like making notes to instantly review them anytime and thus systematically or sequentially reveal the whole picture of an event.
In summary, SkyREC AiMS is dedicated to saving human efforts by effectively digging up the essential threads of a crime through video footage collected from cameras and quickly reporting the route or last-known location of the target objects. Furthermore, by designing the robust foundation of its video analytics algorithm, AiMS supports authorities turn their position from a reactive role to even a proactive solver by enhancing the incident response time and post-event investigation efficiency.
