Rohan Sanil and Jinjun Wang are the founders of Deep North, they founded the company in 2016 to enable businesses to digitize their physical environments. Rohan has 20 years of solid product and business experience and Jinjun Wang has 15 years of research and development experience in multimedia computing, computer vision and pattern recognition.
What is the Far North USP?
Existing video analytics companies use facial recognition to track consumer behavior at physical locations; however, due to privacy restrictions such as GDPR and CCPA, such software cannot be used without consumer consent as it retains personally identifiable information (PII). As a result, most vendors deployed in the West focus on scanning and identifying items from a single vantage point, such as a front door or checkout counter.
Deep North’s main competitive advantage is its re-identification patent, which uses skeleton-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 then stored in our cloud database as an anonymized hash code. We can track individuals across multiple cameras using this algorithm because we can re-identify each skeleton with their encrypted hash code as they move from one camera angle to another, allowing us to to assemble a customer’s journey from arrival to departure in a fully GDPR framework. compliant manner without storing any PII. This type of multi-layered data is much more beneficial for organizations because it allows them to understand the entire customer journey for different cohorts.
Tell us about the product, services and solution that your company offers?
Deep North’s end-to-end software solution combines artificial intelligence and computer vision to help retailers and a business digitize and analyze behavioral metrics in the physical world and gives them tools to act on that information. Deep North’s proprietary entity detection platform enables our clients to gain behavioral insights such as store traffic, queue wait times and conversion through our custom dashboards . Deep North’s cross-camera re-identification and tracking technology allows us to provide insights into the buyer’s journey, such as dominant paths and cross-conversions. We work with a wide range of Global 2000 clients in sectors such as fashion retail, supermarkets and shopping malls.
Where is Artificial Intelligence in retail?
Retail has witnessed underlying changes over the past decade. With the rise of online players, retailers have started talking about embracing the omnichannel playbook and phygital transformation. However, it was more talk and less action, in part because the technology was simply nascent to support the ambitions of brick-and-mortar stores. The pandemic has further exposed the glaring inadequacies of a traditional retail setup to deliver a seamless customer experience. The challenge in a physical environment is indeed multiple and we are entering a phase of retail renaissance fueled by a convergence of technologies.
One of the enablers is artificial intelligence which can provide insights for operational and strategic decisions. With current advancements in AI, especially in the subfields of computer vision and deep learning, retailers can access rich granular insights beyond POS transaction data and counting. traditional door. Businesses can understand the challenges of each of their stores and empower their in-store teams with real-time insights such as steps per day and per hour, dwell times in different areas, POS conversions from zone, alerts to open additional payment lanes, etc. The possibilities are immense for businesses to create custom use cases and this is exactly one area where both challenges and opportunities lie.
How does the Deep North video analytics platform work?
Deep North’s value proposition lies in leveraging existing CCTV cameras. In an on-premises edge-based deployment scenario, video data is processed on-premises to generate rich metadata to detect objects and assign them unique identifiers.
The inference pipeline brings together metadata from camera streams and algorithms for real-time processing in the cloud. The inference pipeline algorithms will generate rich metadata about physical environments such as engagement, pathing, and dwelling. Metadata is then redirected as information to Deep North Platform dashboards and mobile apps with less than a second of latency. When deployed in the cloud, video processing occurs in the cloud. The entire platform is designed to respect privacy, as metadata cannot be linked to PII data.
What markets does Deep North cater to?
Deep North’s customer base spans retail, shopping malls, quick service restaurants, transportation, commercial real estate, manufacturing and warehouses. We help improve operations and create exceptional customer experiences in-store or across the chain using real-time video analytics.
Deep North is currently operational worldwide with deployments in North America, Europe, the Middle East and Southeast Asia. We recently set up our Mumbai office from March this year with a focus on educating businesses in the APAC market on the value that intelligent video analytics can bring to the table.
What is the impact of AI video analytics on business operations
Instead of simply calling IA, we prefer to use the term Intelligent Video Analytics. Because it really accentuates what the technology intends to do. At its core, intelligent video analytics technology gives direct visibility into what’s working and what’s not in the retail space. In a sense, it empowers businesses by giving them full ownership of the physical environment.
Armed with actionable alerts, insights, recommendations and predictive analytics, retailers are able to optimize in-store operations, reduce wait times, prevent cart abandonment, ensure personalized attention to customers in areas of high value-added merchandise, to maximize marketing and promotional impact, etc. The basket of use cases for shopping malls is expanding as they can track customer footfall patterns throughout the day, monitor mall traffic, assess tenant mix health, optimize rental strategies, etc. In short, there is a clear path to ROI that can be continually nurtured and developed in consultation with each business.
What are the future projects of the Far North?
We have big goals to democratize the use of computer vision globally. The efficiency and computational power of deploying computer vision-based systems continues to grow significantly year after year as companies like NVIDIA continue to invest billions in improving GPU microarchitecture. Moreover, if the 5G network infrastructure is implemented on a large scale, it becomes more practical to use an offsite architecture since the video streams from the cameras can be transferred directly to the cloud.
These developments allow Deep North to better serve a wider range of customers as we can easily and cost-effectively light their stores at night. So we are stepping up our attempts to expand our geographic reach outside of the West to include places like the Gulf, India and Southeast Asia.
By implementing our solution with businesses in each new region, we are able to create new use cases from businesses with varying needs and perspectives, while training and improving the accuracy of our algorithms in these new contexts.
A few words about Techiexpert
We are excited to reach industry stalwarts, researchers, tech professionals, enthusiasts across Techiexpert’s wide readership. At Deep North, we see ourselves as an accelerator for the adoption of privacy-compliant computer vision technology in consumer-facing businesses. This requires a concerted effort to educate businesses and customers. We’re all in this together and to assume effectively, platforms like Techiexpert, as part of the broader AI ecosystem, play a pivotal role in advancing the convergence of technologies to create a human experience. rewarding overall.