Importance of accuracy in video analytics March 6, 2016

Thinking about analytics?

Analytics is a real buzz word at the moment. Google analytics, video analytics, web-based analytics. It all sounds exciting and forward-thinking, but what does analytics really mean, and how can businesses be assured that investing in such new technology will actually be of benefit? Today, there are so many tools available that it is almost overwhelming. Data accuracy should be a key consideration when looking into analytics.

Progression of the video analytics space

Video analytics are of particular interest at the moment, especially as offline commerce attempts to compete with the ever-evolving and prevalent online space.

So what is video analytics all about?

A decade ago, video analytics was very much focused around simple surveillance and analysis of situational influences on security. Motion detection and license plate recognition were about the extent of capabilities, and the software was very much cumbersome in terms of data extraction and usability.
Today, video analytics provides real business intelligence, equipping management with valuable data to improve operations, sales, marketing and more. Video analytics now refers more to the pixel tracking of customer movement within a camera’s view, giving an eye into the world of the consumer as they move around and interact with physical stores.

These days, a camera on its own is very much like a traditional mobile phone without the intelligence of a new smart phone. Of course, you would choose the phone with smarts. Smart retailers are making the switch to intelligent video solutions in order to understand customer behaviour within physical stores and improve the overall in-store customer experience.

The accuracy question

Foot traffic data, queue times, heat mapping, dwell analysis and demographics are some of the new and exciting analytics which enable businesses to measure conversion rates, optimise staffing, and improve product placement. This sounds beneficial, but can the integrity of this data be assured and is it accurate enough to generate actionable insights?
Unfortunately, the answer is often no.
Accurate video analytics are dependent on a multitude of factors which are quite often overlooked. Camera field of view, positioning, and lighting are just a few elements that must be considered in order to generate accurate data. Cameras placed near highly variable lighting or which capture background objects, for example, will skew data, and result in less than accurate results.

Finding the best product

Consider asking a few key questions before choosing an analytics provider. This will help you determine the accuracy of data you are receiving, and how much you can rely on the data to provide actionable insights.

1. Are cameras calibrated to ensure environmental factors are not influencing data?
2. Is the analytic based around trending data or pixel tracking? Pixel tracking is significantly more accurate in capturing data.
3. Is the data cloud-checked to verify accuracy?
4. Can you ensure cameras are installed with optimal positioning to capture the correct field of view for accurate analytics data?
5. Do you have a technical support team monitoring cameras to audit for accuracy and reduce the risk of data loss?

AFN works closely with analytics partners to provide Australian businesses with highly advanced and accurate video analytics solutions which answer ‘yes’ to the above questions. Video analytics can play an important role in business decision-making, and enables data gathered from video surveillance to improve and expand a business rather than just protect it.

Related Posts