[Exclusive Interview] Skeletal-based Tracking Algorithms For Video Analytics By This Startup Is A Gamechanger!
Recently we interacted with Mr. Azhaan Merchant, SVP of Strategy & Business Development at Deep North., where we asked him about his company’s vision, and plans for the future.
Here are the highlights:
Brief about your company since its inception?
Rohan Sanil and Jinjun Wang, experts in multimedia signal processing, pattern recognition, computer vision, and analytics, launched Deep North in 2016. Sanil pioneered the object detection technology in 2016 to drive targeted advertising from online video. When a prominent brand noticed this, they pushed him to develop a method for detecting, analysing, and categorizing things caught by security video cameras in theme parks. This investigation sparked the creation of a product that would unlock the potential of CCTV and security video cameras within the customer’s physical area and apply object detection and analysis to any type of footage. Deep North was then founded to fully realize AI’s disruptive potential in the physical world, allowing businesses to provide better, safer, and more remarkable consumer experiences.
Deep North extended the accessibility of its computer vision and video analytics techniques to a range of businesses after rebranding in 2018, including retailers, supermarkets, airports, shopping malls, restaurants, and events.
What is the USP of the Deep North?
Existing video analytics companies employ facial recognition to track a consumer’s behaviour across physical venues; however, due to privacy restrictions such as GDPR and CCPA, such software cannot be used without the consumer’s agreement because it maintains personally identifiable information (PII). As a result, most vendors deployed in the West, concentrate on analysing and identifying items from a single point of view, such as an entrance doorway or a check-out counter.
Deep North’s major competitive advantage is its reidentification patent, which employs skeletal-based tracking algorithms that separate each unique skeleton into 124 distinct vectors. The combination of these vectors, comparable to a fingerprint, is unique to each individual and is subsequently saved in our cloud database as an anonymized hash code. We can track individuals across multiple cameras using this algorithm because we can re-ID each skeletal with their unencrypted hash code when they move from one camera angle to another – allowing us to stitch together a customer’s journey from arrival to depart in a completely GDPR-compliant manner without storing any PII. This kind of multi-layered data is far more beneficial for organizations since it allows them to comprehend the whole client journey for different cohorts.
- How AI helps businesses in winning shoppers’ confidence and turn them into customers?
AI and IoT have begun to push the boundaries of innovation by providing enterprises with access to various data streams. The use of data enables platforms to provide personalized services to users by identifying patterns that help them optimize their services.
Whereas most AI systems collect and report data, the Deep North platform provides organizations with real-time alerts and suggestions to help them remove bottlenecks and enhance business outcomes. Our goal is to provide our customers with a thorough insight of consumer behaviour as it relates to each of their brick-and-mortar stores. Enterprises are using our computer vision technology as an extension of their team, capturing what’s working and what isn’t in real time and then offering rapid, actionable solutions to help their customers have a superior shopping experience. Where should they place certain products? How many sales associates do they need at every hour? Do they need to open up another till to expedite the check-out time?
Businesses who have ingested our insights have seen clear improvements in conversion rates, and have progressively developed trust in our platform to help them optimize their brick and mortar stores.
- What are elements that work to enhance customer Engagement?
Most operators currently only have access to point-of-sale data within their brick-and-mortar stores. As a result, there is a complete black hole of data on what a customer does from the time they enter the store to the time they left. Deep North assists operators in unlocking various new tranches of data relating to what the client is doing at every point – What demographics are most likely to visit the store? Where did they spend their time? What products did they interact with? Having access to this data allows operators to better understand the consumer funnel and optimize the in-store experience in each of their locations.
Deep North’s technology continuously optimizes over time after this data is unlocked, and its predictive analytics engine advises corporations on how they can operate their stores to enhance customer engagement.
- What is the scope of AI for revenue generation and ensuring repeat customers?
Deep North’s computer vision and AI-driven decision-making helps store operators boost productivity and profitability by ensuring they minimize lost revenue opportunities. Few of the key reasons shoppers abandon a purchase are due to reasons such as out of stock shelves, no assistance from sales associates, incorrect product mixes or long lines at the check-out.
Using AI such as computer vision, operators can understand when, where and how often their customers are facing these problems. Additionally, they can understand this in real-time across hundreds of stores or millions of square feet of shopping floors. Once they have access to these kind of in-store analytics at scale, they can diagnose what the systemic problems are within their organization and implement change on the floor. This will ensure customers have better shopping experiences and stay loyal to the brand.
- Explain how it is helpful in creating a safe work environment?
Workplace safety is and should always be an organization’s top priority. Deep North’s computer vision solution is deployed in large warehouses such as fulfillment and distribution centers where we utilize the existing camera infrastructure to monitor the movement of vehicle, people and machinery with the goal of improving worker safety.
Our use cases range from detecting if workers are wearing their helmets/ high-vis jackets, to seeing if vehicle operators are driving within compliant paths, to providing real-time alerts if we detect a potential hazard.
- What are the future plans of the Deep North?
We have ambitious plans to democratize the use of computer vision across the world. As firms like NVIDIA continue to invest billions in upgrading GPU microarchitecture, the efficiency and computational capability of deploying computer vision-based platforms improves tremendously year after year. Furthermore, once 5G network infrastructure is deployed at scale, it becomes more viable to employ off-premise architecture since video footage can be uploaded straight to the cloud from cameras. Owing to these advancements, Deep North can work with a larger number of clients as we will be able to light up their stores overnight with minimal cost and difficulty.
As a result, we are doubling our efforts to expand our geographic presence from the West to areas like the Gulf, India and Southeast Asia. Deploying our solution with enterprises in each new geography not only helps us train and improve the accuracy of algorithms in new environments but it also allows to develop new use cases from enterprises with different perspectives and requirements.