Interviewed By Akhlaqur Rahman Sachee, Team MBR

Sigmind is a pioneering AI-focused startup specialising in advanced computer vision technologies, founded in 2017 with a mission to revolutionise public safety, traffic management, and security through innovative solutions. Leveraging deep learning and artificial intelligence, Sigmind transforms visual data from surveillance cameras into actionable insights, enhancing the operational efficiency of enterprises and governments. With its flagship products like TrafficFlow and Watchcam, Sigmind offers a comprehensive suite for real-time traffic and human analytics, setting new standards in surveillance automation. Team MBR was in a conversation with the founder and CEO of Sigmind, Mr. Md. Abu Anas Ibn Samad, and had the opportunity to learn about his inspirations behind forming the startup and his future aspirations surrounding Sigmind.

Akhlaqur Rahman Sachee: Sigmind has recently created a buzz on social media with the introduction of TrafficFlow, a solution that employs video analytics for real-time traffic management. Would you kindly share with us how you came up with the idea to form Sigmind?

Md. Abu Anas Ibn Samad: The genesis of Sigmind was driven by a vision to harness the transformative power of artificial intelligence and computer vision to address some of the most pressing challenges faced by modern societies. Our journey began with a keen observation of the burgeoning challenges in urban management, particularly in the context of Bangladesh. Traffic congestion, a lack of efficient public safety measures, and inadequate security infrastructure are not just inconveniences but significant barriers to economic growth and societal well-being. These issues are compounded by the rapid urbanisation and growth of digital infrastructure, creating a pressing need for innovative solutions.

Dhaka is the slowest city, according to the US-based National Bureau of Economic Research. A study by BRTA shows that traffic congestion costs Dhaka approximately USD 3.2 billion annually in lost productivity and fuel waste. The founding team members of Sigmind, equipped with backgrounds in AI, machine learning, and software engineering, recognised the untapped potential of AI-powered computer vision technologies to solve this issue. The idea for TrafficFlow, our flagship solution, is a direct response to the chaotic traffic conditions in Dhaka. We envisioned a system that could not only monitor but also analyse and predict traffic patterns, offering actionable insights for traffic management. By employing state-of-the-art video analytics and deep learning algorithms, TrafficFlow has been conceptualised as a solution that could revolutionise traffic management by providing real-time data, enhancing decision-making, and ultimately improving the quality of urban life.

The positive reception of TrafficFlow on social media and beyond is a testament to the relevance and impact of our solutions. The formation of Sigmind was underpinned by a commitment to innovation, excellence, and social impact. As we continue to innovate and expand our offerings, the foundational idea of using technology to solve real-world problems remains at the heart of Sigmind.

Akhlaqur Rahman Sachee: As video analytics facilitates the study of object attributes, movement patterns, behaviours, and so on, the scope of applications of this technology is quite wide. May we know what areas Sigmind has already explored and what areas it intends to explore in the near future?

Md. Abu Anas Ibn Samad: Sigmind has harnessed the capabilities of video analytics to address a broad spectrum of applications, focusing on enhancing public safety, optimising traffic management, and improving security measures. Our exploration of the potential of video analytics has led to the development of innovative solutions that cater to diverse needs. Below are the areas we have explored.

Traffic Management: With our flagship product, TrafficFlow, we have made significant strides in real-time traffic management, offering solutions for vehicle detection, classification, speed estimation, and congestion analysis to improve traffic flow and reduce congestion.

Public Safety: Our solutions contribute to public safety by enabling crowd management, detecting unauthorised gatherings, and identifying potential safety hazards in real-time. This is crucial for event management, urban planning, and emergency response coordination.

Security and Surveillance: We've applied video analytics to enhance security measures through facial recognition, intrusion detection, activity monitoring, and attendance monitoring. These capabilities are vital for securing public spaces, critical infrastructure, and private properties.

Retail Analytics: Our technology also finds application in the retail sector, where it can track customer movement, analyse behaviour patterns, and optimise store layouts to improve the shopping experience and increase sales.

Industrial Automation: We have started to explore the use of video analytics for industrial automation, including monitoring equipment performance, ensuring safety protocols are followed, and optimizing manufacturing processes.

We have plans to explore the following areas in the near future.

Smart Cities: We intend to deepen our involvement in smart city initiatives by developing solutions that integrate with various urban systems for enhanced city-wide management, energy conservation, and sustainable urban development.

Education: Enhancing security on campuses and optimising educational resources through behaviour analysis and space utilisation are areas we see significant potential in.

Agriculture: Leveraging video analytics for precision agriculture to monitor crop health, detect pest infestations, and automate harvesting processes is another avenue we're excited about exploring.

Environmental Monitoring: We are keen on developing solutions for environmental monitoring, such as tracking wildlife populations, detecting forest fires early, and monitoring air and water quality, to contribute to conservation efforts.

In addition to the aforementioned areas, Sigmind is keenly exploring the integration of large language models (LLMs) and generative AI into our video analytics solutions. This innovative approach aims to enhance predictive video analytics and forecasting capabilities, offering even more sophisticated insights into security, traffic management, environmental monitoring, and more.

Akhlaqur Rahman Sachee: Sigmind has already provided vehicle analytics and human analytics solutions to BEPZA, PMO, CAAB, BCB, Finance Division, Robi Axiata, and some other local organisations. Would you kindly share with us the business model of Sigmind and how it makes money out of the solutions it provides?

Md. Abu Anas Ibn Samad: Sigmind operates under a business model that is designed to offer scalable, cutting-edge AI-powered video analytics solutions to a diverse clientele, including government bodies, enterprises, and private organisations. Our revenue generation strategies are built around providing high-value, customisable solutions that cater to the specific needs of our clients, ensuring a high level of satisfaction and return on investment. Here is an overview of how Sigmind makes money from its solutions.

Sales of Software Licenses and Hardware Systems: We sell licenses for our proprietary video analytics software, including TrafficFlow and other human and vehicle analytics solutions. These licenses can be offered on a perpetual basis or through annual renewals, depending on client preferences. For clients requiring hardware, we provide integrated solutions that combine our software with specialised hardware, such as GPU-accelerated servers or embedded systems. This ensures optimal performance of our analytics solutions.

Service Subscriptions: Sigmind also adopts a software-as-a-service (SaaS) model for clients preferring to use our solutions without a significant upfront investment in hardware and software licenses. This model provides access to our video analytics platform for a recurring subscription fee, making it flexible and scalable for various business sizes. Leveraging cloud computing, we also offer cloud-based video analytics services that allow clients to analyse video data without maintaining their own computing infrastructure.

Custom Solutions and Consultancy: We engage in custom solution development for clients with unique requirements. This includes tailoring our existing software, developing new algorithms, or integrating our solutions with existing systems. Alongside, Sigmind provides consultancy services in AI and computer vision, helping clients design and implement effective surveillance, security, and traffic management strategies. This consultancy extends to system design, deployment strategies, and optimisation of video analytics operations.

Maintenance and Support Services: For ongoing support and maintenance of our systems, we offer annual maintenance contracts. These contracts ensure that clients receive software updates, system maintenance, and technical support, providing a steady revenue stream for Sigmind. We also offer comprehensive training programmes for clients’ technical and operational staff, ensuring they can effectively use and maintain the video analytics systems.

Partnerships and Collaborations: Collaborating with hardware vendors, software developers, and service providers expands our market reach and opens up new revenue streams through joint ventures and co-marketing efforts. By securing contracts with government agencies and large enterprises for city-wide surveillance, traffic management, and public safety projects, we ensure long-term engagements that contribute significantly to our revenue.

Akhlaqur Rahman Sachee: Solutions that employ video analytics can be inappropriately used, breaching the privacy of others. How does Sigmind ensure that its solutions are used for the purposes for which they are intended to be used?

Md. Abu Anas Ibn Samad: Sigmind is aware of the ethical considerations and privacy concerns surrounding the use of video analytics and facial recognition technologies. We are committed to ensuring that our solutions are employed responsibly and for their intended purposes, which are to enhance public safety, streamline traffic management, and bolster security measures without infringing on individual privacy rights. To this end, we have implemented several key measures.

Privacy by Design: Our systems are developed with privacy as a core principle. We employ techniques such as data anonymization and encryption to protect individuals' identities. Our solutions process visual data in a way that emphasises the privacy of individuals, ensuring that personal data is not unnecessarily exposed or stored.

Compliance with Regulations: Sigmind strictly adheres to local and international data protection regulations, including GDPR in applicable regions. We work closely with legal experts to ensure our products comply with all relevant laws and guidelines, particularly those governing privacy and data protection.

Ethical Guidelines: We have established ethical guidelines for the deployment of our technologies. These guidelines mandate that our solutions are used in a manner that respects human rights and privacy. We conduct thorough ethical reviews of potential deployments, rejecting applications that pose a significant risk of misuse or harm.

Client Education: Sigmind actively educates clients on the responsible use of video analytics technology. We provide guidance on best practices for protecting privacy, including recommended operational protocols that minimise the risk of privacy violations.

Transparency and Accountability: We advocate for transparency in the use of video analytics and facial recognition technologies. Our policies and practices are openly communicated to stakeholders, and we hold ourselves accountable for the responsible deployment of our technologies.

Selective Deployment: Sigmind is selective about the projects and clients we work with, choosing partners who share our commitment to ethical practices and who intend to use our solutions for public good, such as enhancing safety and security in a non-invasive, privacy-respecting manner.

Through these measures, Sigmind aims to lead by example in the responsible development and application of AI-powered video analytics, ensuring that our innovative technologies contribute positively to society without compromising the privacy and rights of individuals.

Akhlaqur Rahman Sachee: Would you kindly share with us the deployment procedure for the video analytics solution that Sigmind provides? Would the existing surveillance setups of the users be sufficient to integrate Sigmind’s solutions, or would the users need to replace the existing setups and purchase customised hardware from Sigmind?

Md. Abu Anas Ibn Samad: Sigmind’s deployment procedure for our video analytics solutions is designed to be seamless and efficient, ensuring minimal disruption while maximising the utility of existing infrastructure. We start by understanding the specific requirements, challenges, and objectives of the client. This involves assessing the client's current surveillance and security setup and identifying areas where our video analytics solutions can add value. A technical assessment of the existing surveillance infrastructure is conducted to determine compatibility with our solutions. This includes evaluating the specifications of existing cameras, network capabilities, and hardware to ensure our software can be integrated effectively. It is to be noted that Sigmind’s solutions are designed to be highly compatible with a wide range of IP cameras and surveillance systems. Our software can often be integrated with existing setups without the need for extensive hardware upgrades, leveraging the current investment to its fullest potential. For scenarios where the existing infrastructure cannot support advanced analytics or where enhanced capabilities are desired, Sigmind offers customised hardware solutions. These include GPU-accelerated embedded systems and servers that are optimised for high-performance computing needs in video analytics.

Our technical team oversees the integration of Sigmind’s software with the client's surveillance system, ensuring seamless operation. If new hardware is required, we assist in its installation and configuration to work in tandem with existing setups. The system undergoes rigorous testing and calibration to ensure accurate analytics, optimal performance, and reliability. This phase includes tuning the algorithms to the specific use-case scenarios of the client, such as specific traffic patterns or security monitoring needs.

Sigmind provides comprehensive training to the client's staff, covering system operation, data interpretation, and maintenance procedures. This ensures that the client can effectively utilise the system from day one. We also offer ongoing technical support and system updates to ensure the analytics solutions continue to perform optimally and evolve with the client’s needs. This includes software updates for improved features and capabilities.

Our solutions are designed with scalability in mind, allowing for easy expansion or upgrades as the client's requirements grow or change. This includes adding more camera inputs, enhancing analytics capabilities, or integrating new AI models.

Akhlaqur Rahman Sachee: Video analytics may fail to detect objects with accuracy and generate misleading data. Would you kindly shed some light on the statistics regarding the accuracy of TrafficFlow, which is currently trending?

Md. Abu Anas Ibn Samad: TrafficFlow, Sigmind’s flagship video analytics solution designed for traffic management and vehicle analytics, represents the forefront of our AI-driven technology suite. The effectiveness and reliability of TrafficFlow are grounded in its sophisticated algorithmic foundation and extensive validation across diverse real-world scenarios. TrafficFlow achieves a high degree of accuracy in vehicle detection and classification, distinguishing between 25 distinct vehicle categories specific to the Bangladeshi context. The system consistently demonstrates over 98% accuracy in vehicle detection and classification under varied lighting and weather conditions. The licence plate recognition feature of TrafficFlow, crucial for vehicle identification and tracking, shows an accuracy rate of over 99% in broad daylight, while it falls to nearly 95% in low-light conditions. This is particularly notable given the challenges of diverse plate designs, varying conditions of visibility, angles of capture, and lighting conditions. With our innovative recurrence sorting method, we are able to achieve state-of-the-art licence plate recognition from a broader camera perspective. For measuring vehicle speed and direction, TrafficFlow employs optical flow-based algorithms, providing accurate estimations that are critical for traffic management applications. The system achieves an accuracy rate of approximately 91% in speed estimation and direction detection.

However, Sigmind is committed to continuous research and development to further enhance the accuracy and performance of TrafficFlow. Our AI models are regularly updated with new data, improving their ability to adapt to changing conditions and new challenges.

Akhlaqur Rahman Sachee: Sigmind has provided its solutions to some government bodies and some local corporations so far. May we know if Sigmind is interested in providing its surveillance and security solutions for small businesses and private properties? If positive, what pricing plan does it provide for small-scale subscribers to accommodate their needs?

Md. Abu Anas Ibn Samad: Sigmind is dedicated to democratizing access to advanced AI-powered surveillance and security solutions, not only for government bodies and large corporations but also for small businesses and private property owners. Recognising the unique challenges and budget constraints that smaller entities might face, Sigmind has tailored offerings designed to bring the benefits of cutting-edge computer vision technologies within their reach. Our approach ensures that smaller entities can benefit from high-quality surveillance without the need for extensive infrastructure or significant upfront investment.

Our pricing strategies are designed to accommodate varying levels of usage, ensuring that small businesses and private properties can select a plan that best suits their security needs without overspending. We offer subscription-based services that allow smaller entities to access our surveillance and security solutions without a hefty initial investment. These subscriptions can include regular updates and maintenance, ensuring that clients always have access to the latest features and improvements. In addition to that, our solutions are modular, allowing clients to choose and pay for only the features they need. Whether it is basic surveillance, vehicle analytics, or facial recognition capabilities, clients can tailor their package to match their requirements and budget. To help small businesses and private property owners understand the value of our solutions, we offer trial periods and demonstrations. This allows potential clients to experience firsthand the benefits of our technology before making a commitment.

All our clients, regardless of size, receive comprehensive support to ensure the smooth operation of their surveillance systems. This includes technical support, software updates, and training on system usage and maintenance. We provide custom training sessions tailored to the specific operational needs of small businesses and private properties, ensuring that staff can effectively manage and utilise the surveillance system.

Akhlaqur Rahman Sachee: The solutions Sigmind provides are less likely to face geographical barriers, or the solutions may be useful in other geographical boundaries with slight tweaking. Would you kindly share with us if Sigmind has explored cross-border business opportunities?

Md. Abu Anas Ibn Samad: Sigmind recognises the universal applicability and potential of its AI-powered surveillance and security solutions to transcend geographical boundaries. The core technology behind our offerings, grounded in advanced computer vision and artificial intelligence, is highly adaptable and can address similar challenges faced by entities worldwide. Given the global nature of security and surveillance needs, our solutions are designed to be customised to various environments and legal frameworks, making cross-border applications not only feasible but highly promising.

Sigmind has conducted comprehensive market assessments to understand the specific needs, challenges, and regulatory landscapes of potential international markets. In addition to its impactful presence within Bangladesh, Sigmind has broadened its horizons by exporting its cutting-edge software solutions to several countries around the globe, including the United Kingdom, the Netherlands, Japan, India, and others. This expansion is a testament to the universal applicability and high demand for advanced AI-powered video analytics technologies across various sectors worldwide.

We have initiated pilot projects in selected countries to demonstrate the adaptability and effectiveness of our solutions in diverse settings. These pilots serve as a foundation for establishing partnerships with local entities and understanding the operational nuances in different geographical contexts. An essential part of our cross-border strategy is ensuring that our solutions comply with international data protection and privacy standards, such as GDPR in Europe. This commitment to compliance facilitates smoother entry into international markets by aligning with local regulations. Recognising the importance of local context, from traffic patterns to architectural differences in surveillance environments, our team works closely with local partners to customise our solutions. This includes adjusting models by fine-tuning them to local demographics, vehicle types, and even specific security threats.