Video analytics is costly and open to security and privacy risks, so be sure it's right for your company before implementation.
Video has established its role as an essential and mission-critical technology during the COVID-19 crisis. Its role has been substantiated with every video conference, virtual performance, tele-health appointment, and tele-learning session.
The expanding role of video in business has also paved the way for more big data analytics that operate on video and photographic data—with an end goal of delivering actionable, visual intelligence.
SEE: How to Choose the Right Video Analytics for Business, Operational, or Security Intelligence (TechRepublic download)
Among the industry use cases for video are hospital asset tracking; building, workspace and warehouse usage monitoring; thermal imagery for fever detection; methane gas leak monitoring; and health monitoring and counting of livestock.
"From this video analysis, you can tell what people in crowds are doing, develop intelligence about worker productivity, and build a record of intelligence such as how much time is needed to complete a given task," said Phil Ressler, CEO of Sixgill, which provides decision automation for visual data. "In a senior living facility, you can observe whether resident behavior patterns have changed, which could indicate dementia."
Using video and developing artificial intelligence (AI) and machine learning automation to assist with decision making can render times to decision faster and more impactful, provided organizations adhere to best practices and privacy standards.
"For organizations considering an expanded use of video analytics, a first step is to identify a cluster of problems they want to solve with video," Ressler said. "I strongly encourage them to use a cluster of problems approach in developing a robust business case—and to not just settle on one app they want to do. The ideal number is to identify three different initial projects for sensor networks and video."
SEE: Special report: Managing AI and ML in the enterprise (free PDF) (TechRepublic)
Ressler also said the biggest inhibitor for companies launching video analytics projects is a limited network of available integrators. "We do many of the integrations ourselves for this reason," he said.
For companies looking to expand their use of video and video analytics, here are three recommendations:
Consider a broader business picture
Ressler suggested this could be done by identifying a broad range of business cases instead of just one for video analytics. While this might fly in the face of what historical analytics wisdom has been ("start small with one business case"), it can be a viable strategy with video because investments in video, networks, and bandwidth can be substantial. You want to be able to leverage these assets for as many applications as you can in the present and in the future.
The recent Zoom privacy debacle during the COVID-19 crisis is an example.
SEE: Zoom adds data center routing, security updates (ZDNet)
The message is clear: While business use cases and purposes widely vary, organizations should develop their privacy policies along with the technology they plan to deploy.
Use automation, but with human checkpoints
Automation and machine learning are great tools for video and analytics, but they aren't foolproof. For critical judgments, human experts should be a second line of review to ensure that the conclusions drawn from the data are accurate.
If you follow these cautions when beginning your video analytics journey, you can look forward to excellent and insightful outcomes taking your company to the next level of business knowledge.
- IT leader's guide to deep learning (TechRepublic)
- Telemedicine, AI, and deep learning are revolutionizing healthcare (TechRepublic download)
- Artificial intelligence ethics policy (TechRepublic Premium)
- What is AI? Everything you need to know about Artificial Intelligence (ZDNet)
- 6 ways to delete yourself from the internet (CNET)
- Artificial Intelligence: More must-read coverage (TechRepublic on Flipboard)