Intelligent Computer Vision Technology for Video Analytics

6 billion hours of footage generated in a day. That is a lot of visual data to process. But these data will not make any sense if we take thousands of hours to analysis and we put huge work force to operate these cameras. Current security system does not have any real time crime control notification and there are no real time search options. Sigmind use machine learning to analyze video data. Our technology can be used to Face ID, recognize License plates, vehicles and is effectively used for Security & surveillance, Traffic Management Systems, Attendance Tracking, Foot traffic and much more. Sigmind provides solution for people who want to automate CCTV monitoring and Post Footage Analysis. The system works perfectly with the client's existing CCTV/IP cameras, replacing only the NVR/DVR box with its A.I driven unique intelligent video surveillance system. This could drastically change the security scenario in Bangladesh.

Current security camera lacks lighting and pose independent facial recognition, lacks Suspicious activity detection and instant notification, no real time crime control notification, no real time search options, requires large work force and there is no object (bag,suitcase,gun) detection. Moreover, due to the lack of instant incident detection and crime reporting features, those cameras could not prevent crime from happening. They only record the footage. Currently there is no Bangladeshi vehicle (micro bus, mini bus, heavy truck, mini truck....) detection system, no Bangla license plate identification system. In case of the house owner, the cameras installed at their premises cannot prevent theft or intruder. However there is also no foot traffic analytics in retail shops to address customer measurement, demography and persona based targeting. The price of A.I camera is still expensive. Sigmind addresses these problems and clients and proposed a unique video analytics solution that can serve the national law enforcement agencies, house owners, parking and toll plaza management, retail shops/enterprises and whoever wants to automate the cctv cameras in a cost effective way to extract valuable video data.

Sigmind aims to serve the law enforcement agencies to detect and identify people in real time or from a pre-recorded video footage to help them accelerate faster response to a crime and extract valuable data in almost no time. For house owner, our solution is used as a smart monitoring system to prevent intruder from gaining access to the house or instant notification in case of theft or any unusual indecent. In the area of traffic management, our video analytics serve to read Bangla license plates along with identification of 14 categories Bangladeshi vehicles. It also can detect lane violation. For parking management, we automate the conventional method by detecting vehicle types and number plates as well as empty slots in parking zone. In toll plaza, Sigmind counts different types of vehicles automatically and reduce the work force to save huge amount of money and hassle. In retail we can track customer journey in store leading to greater insights to help improve their sales.     

Sigmind integrated state-of-the-art machine learning based technologies to detect persons, objects, faces, firearms in real-time from multiple camera feeds on its powerful server.

Sigmind figured out the gap between giant companies and need of Bangladesh. So Sigmind developed a localized product that can detect local Bangladeshi vehicles like mid truck, big truck, mini bus, micro bus (even the modified one) along with 14 other types. Moreover Sigmind can identify Bangla License Plates. One of the biggest draw back for facial recognition is, it is good to detect white faces whereas Sigmind has developed its training model and CNN such a way that it is as accurate detecting south asian faces as white ones. These are the innovations that Sigmind claims. 

We have calculated the time takes a person to analyze a single 24 hours video footage. Almost 4/5 hours. Sigmind analyses hours of footage in almost no time. An ordinary toll plaza requires 10-12 people to collect tolls & counting vehicles. With Sigmind only 2 cameras can do the job. Moreover we human are prone to error. Almost 80% of information is missed from a human eye while operating several cameras but the Computer misses nothing. However our toll plaza clients thanks us that we saved their system loss worth 4 lakh taka per month. Most importantly since current video surveilance system has no real time feed back so it only records footage. While Sigmind can give instant notification when an unusual incedent takes place. 

1. The current number of people we are serving: 6

2. The number we will be serving in one year: 30

3. The number we’ll be serving in five years: 4000+

Within 2019, we would like to make the product more market fit. Our goal for this year to saturate the product in the market, increase our managerial capabilities and of course bringing more people under our umbrella. Within the next 5 years we would like to see ourselves competing in the world market, exploring new dimensions of computer vision technology and bringing more innovations that will have a meaningful effect to this part of the world.

The current barrier is man power. With an insufficient number of team members, we are already overwhelmed with charges. Moreover to train the AI model we need huge amount of data. Collecting, sorting and labeling is a time consuming process. We require google cloud or amazon cloud to get the process faster. 

Competing with giants can be a great barrier for the next five years

We have requested to google authority to permit us use google clouds. If that doesn't help, we also have access to GEN accelerator program. Perhaps we will get help from there. We are currently on process to hire more people to make a larger team.

We think the best for a startup is not competing with the giants. We expect Mergers & Acquisitions

4 core team members 

1 part time staff

3 full time staffs 

 Our strength is our team. We have very clear understanding between us and we know exactly what we are doing. With this same team we have come to this far, from a concept to a proto-type to the market ready product. And we have done it in just 1 year. This was possible only because of the dedication of the members. We have shown our competitiveness in various national and international arenas. This is the reason i think we will be capable to taking any challenge.
Tanvir is the co-founder and serves as an embedded engineer in Sigmind. He has got his degree from Germany in electrical engineering and IT, has worked in Continental automotive GmbH and Airbus defense and space, Germany. Abu Anas is a mechanical engineer from Bangladesh university of engineering and Technology serving as lead programmer. Kazi graduated from Minnesota State university in economics and he is responsible for the business plan, branding, promoting. Arif, a BBA degree holder serving the team as a organizer, handling legal matters and as a CXO. 

Sigmind goes for direct sales with the product that is new to market. Focusing on end users, Sigmind developed 6 SigBox that can handle from 4 to 500 cameras depending on client's requirements. Sigmind has two business models. B2C and B2B.  In B2C market, Sigmind sales its 4 and 50 camera system to house owner, toll plaza management and enterprises to track employee/student attendance.

Video analytics market is huge, 62.7 billion to be exact and there is a market demand of this kind of product. We are focusing to meet our upcoming clients needs and making the product more customizable. So undoubtedly we will achieve financial sustainability through sales. However, in the mean time we are looking for investors to raise capital. 

Tiger IT foundations can help us with its networks to expand in new markets. Of course attracting investors is one of the goal for a startup. So we also would like to attract investors and here Tiger IT can play an important role as a mentor.