Traditionally, dominated by legacy analog systems, India video analytics market is set to take the next leap as AI powered learning systems rapidly become the industry standard, and IP cameras penetrate the market through various national and international hardware vendors. This perfect storm is driving the demand for analytics and surveillance software, solving various types of real-life problems in different industrial segments of the country. But while the allure of this tremendous potential is true, India also poses a unique set of challenges for companies specializing technology-based solutions for the manufacturing and heavy engineering sector. Some of these factors, make usable computer vision solutions, very tricky to build.

For example, in India, as a computer vision based solutioning company, we routinely need to consider unpredictable power fluctuations while building continuous monitoring solutions, and budget for additional power backup safeguards. Or given the natural diversity and extreme nature of weather and climate in some areas, harsh environmental conditions often overheat cameras left in the open, to 50 degrees or more or depending on the location, or plummet the operating temperature close to 0 degrees or lower. Both cases make cameras prone to mechanical malfunctions. Unless planned for in advance, these factors tend to lower the efficacy of our solution in the open market.

In another reminder of expect the unexpected, we have to carefully evaluate the data collection mechanism first, before accepting any client contract. As we all know, for training algorithms to succeed, high quality data collection is critical, but it often becomes impossible to achieve, due to various non-standard practices in the Indian unorganized sector. Examples could include Community-Based Participatory Research (CBPR) and Oral Histories as data collection methods and various non-standard processes dominating small batch production, artisanal manufacturing, custom fabrication areas. For effective solutioning, we need to carefully evaluate first, if the integrity of the collected data from our customer sites, will actually help in powering our vision-based algorithms. And even when everything works as planned, the usability of any solution is still uncertain, as it still depends on the average worker’s ability to undergo training, interaction with new technologies and interpreting their outputs accurately, in certain cases, this can pose as a big challenge in the post implementation phase of certain projects.

Some other things to consider in a diverse country like India are cultural factors, privacy concerns, lack of understanding and resistance to technological change among workers etc. But one factor that we have encountered often is one of sensitization. It requires careful mitigation right at the beginning of any project, and it takes the form of convincing the decision-makers that the same CCTV footage unused until something untoward happens, can actually be used proactively to increase the safety, security and productivity of their business. Sometimes more than the technical challenges, building a case for management support becomes the hardest hurdle to overcome for a AI based solutioning company like Starinco.

Starinco brings sector specific AI enabled computer vision solutions. The starinco team comprises of data scientists & subject matter experts enabling industry soecific visual analytics use cases.

Our learning is, expect the unexpected, and consider ramification of all environmental, technical and human and factors carefully, while accepting a video analytics project in India, finding how to technilly solve problems comes much later.