Evolving Landscape Of AI driven Video Analytics The technology is enabling and empowering COVID-19 management across segments and industries

The Internet has an ocean of images and videos that exude diverse meaning/context to its creator. It can become an extremely informative/relevant source of information should there be a way this data can be read, analysed to arrive at coherent information, and applied to derive meaningful applications, beneficial to a larger audience. Video surveillance is a common technology owned by public and private entities alike to monitor their survillence requirements. Video analytics on the other hand is an advanced emerging technology that gathers, processes and analyses the visual content from the CCTV/IP cameras, making the data an asset that can be put to meaningful use to empower businesses.

Nazi-occupied Germany was one of the first countries to possess video surveillance to observe the launch of long-range guided ballistic missiles. Later, the first video surveillance system was installed in 1942 for observing the launch of V-2 rockets (long-range ballistic missiles) but later on, evolved into other industrial use cases. Cut to the 1970s the evolution of video analytics began with VCR technology making it easier for companies to record video camera feeds. However, this offering only facilitated the storage of data without the ability to read or derive any useful information from it.

The technology witnessed massive advancements from the 90s, when video surveillance was adopted by retail stores, service centers, warehouses, etc ensuring safety and security, gathering information to boost sales, track inventory, and improve service. Today, video analytics has grown by leaps and bounds to imbibe the latest advancements such as artificial intelligence (AI) and machine learning (ML). The technology has further boosted video analytics by offering a powerful and scalable method for extracting detailed and relevant information from these virtual vision systems.

Video analytics also known as computer vision majorly follows pattern recognition, a technique by way of training a computer model to understand/ read visual data by feeding the image dataset and then subject this dataset to various AI/ML techniques, or algorithms. This process enables the system to derive intended patterns related to pre-defined labels establishing a context from the new images it "sees' from the real-time inputs from camera feeds or media files. Video analytics when applied to these captured media feeds enable the users to see inside the via slicing the feed into frames as images, identify and recognize objects inside those images such as animals, objects (non-living things), colours, and even the texts embedded inside the image which helps establish a logical connection/context concerning the content.

Video analytics when coupled with more evolved technologies in machine learning offers a plethora of criteria that can be applied to the media feeds/ images. Deep learning has been a game-changer in the approach to video analytics. Deep learning aids problem solving by being able to extract common patterns and transforming them into a mathematical equation thereby helping in classifying future pieces of information. Deep learning is a rich methodology, encompassing neural networks, hierarchical probabilistic models, and diverse unsupervised and supervised feature learning algorithms. The recent surge of interest in deep learning methods is testimony to the fact that they outperform all previous state-of-the-art techniques as well as the abundance of complex data from different sources e.g., visual, audio, medical, social, and sensor. With such remarkable progress made by the segment/ offering, this technology thus finds its place across verticals/ sectors and applications.

In the current situation, the potential of video analytics can be well utilised by industries such as healthcare, education, manufacturing, banking, civic services, hospitality, among many others. For instance, in healthcare, video analytics can be efficiently utilised to monitor the movement and access of bonafide healthcare professionals. Similarly, in the manufacturing sector, it can be used to manage workers' movement in an area while addressing the COVID: safety and hygiene protocols. The list of video analytics applications and use cases is endless as the adoption continues and forays in solving complex and impactful problems across multiple domains ranging from safety, security to workforce productivity

Furthermore, public safety is a non-negotiable agenda for any government, enterprises, and utility/service providers managing large crowded public places. Surveillance refers to the processes of focusing systematic and routine attention on certain human behaviours for influencing, managing, protecting, or directing purposes. Video Analytics systems for public surveillance have developed significantly in terms of object detection, tracking, classification, behaviour analysis and optical character recognition for ANPR (Automated Number Plate Recognition). Facial recognition algorithms can detect, recognize a particular person, vehicle thus improving accuracy and reliability to monitor real-time situations, vehicles/persons involved in incidents, or even traffic monitoring for vehicular count and flow and facilitating greater mobility management gathering evidence for law enforcement.

Thus, in a nutshell, the inclusion of video analytics has created endless possibilities and avenues for diverse audiences by creating a significant value proposition. The technology which was accessible to CAPEX heavy public bodies and enterprises is now at offer as a service with minimal infrastructure modifications. Today the service is also available as an end-to-end offering, with hardare, software and services bundled together as a single subscription thus significantly reducing the burden of multi-vendor, multi-technology management and cost-effectively delivering desired results as guaranteed outcomes. With the ever-dynamic nature of technologies and its resulting upgrades, there is immense potential that will chalk inroads into many unchartered territories, assisting professionals to make informed/ aided decisions faster while providing/ enabling enriched and powerful consumer experiences in the future. So, are we there yet? Yes, but partially.

The technology has evolved by leaps and bounds, though its adoption needs to be welcomed at scale and that, in turn, will enable the entire ecosystem to achieve the full potential of video analytics as a service.