Question:

I work in the security department of a European optical retail chain with over 300 stores. We are facing internal theft issues and need to review our video surveillance footage from the past year. This is a time-consuming and tedious task, so I am looking for a video analytics software solution that can help us identify and track suspicious employees based on their clothing and appearance. I also need to know the cost per channel for such a software, as we have multiple cameras in each store. I have heard of some vendors like BriefCam that offer this functionality, but I want to get some expert opinions on the best options and the limitations of video analytics for retail security. Thank you for your help.

Answer:

How to use video analytics for retail security

Retail security is a challenging and complex issue, especially for large chains with hundreds of stores and thousands of employees. Internal theft, or shrinkage, is one of the major sources of loss for retailers, accounting for an estimated 35.7% of the total retail inventory shrinkage in Europe in 2019. To prevent and detect internal theft, retailers need to monitor their video surveillance footage and identify any suspicious behavior or activity by their staff. However, this is a time-consuming and tedious task, as it requires reviewing hours of footage from multiple cameras in each store.

Fortunately, there is a solution that can make this task easier and more efficient: video analytics. Video analytics is the process of applying artificial intelligence (AI) and computer vision techniques to video data, to extract meaningful information and insights. Video analytics can help retailers with various aspects of security, such as:

– Face recognition: Video analytics can recognize and verify the identity of individuals in the video, such as employees, customers, or intruders. This can help retailers to enforce access control, track attendance, and spot unauthorized or fraudulent entries.

– Object detection: Video analytics can detect and classify objects in the video, such as products, cash, or weapons. This can help retailers to monitor inventory levels, detect theft or vandalism, and alert security staff in case of emergencies.

– Behavior analysis: Video analytics can analyze and understand the behavior and actions of individuals in the video, such as walking, running, or fighting. This can help retailers to identify and track suspicious employees based on their clothing and appearance, as well as detect abnormal or violent behavior that may pose a threat to safety or security.

– Scene understanding: Video analytics can understand and interpret the context and situation of the video, such as the location, time, or event. This can help retailers to optimize store layout, traffic flow, and customer service, as well as provide insights into customer preferences, habits, and satisfaction.

There are many vendors that offer video analytics software solutions for retail security, such as BriefCam, Axis, Genetec, Verint, and Avigilon. These solutions vary in terms of features, performance, accuracy, and cost. To choose the best option for your needs, you should consider the following factors:

– Compatibility: The video analytics software should be compatible with your existing video surveillance system, such as the cameras, recorders, and network. You should also check the hardware and software requirements, such as the CPU, GPU, RAM, and OS, to ensure that your system can support the video analytics software.

– Scalability: The video analytics software should be scalable and flexible, to accommodate the growth and changes of your business. You should be able to add or remove cameras, stores, or features, as well as adjust the settings and parameters, without compromising the quality or performance of the video analytics software.

– Usability: The video analytics software should be user-friendly and intuitive, to enable easy installation, configuration, and operation. You should also have access to clear and comprehensive documentation, training, and support, to help you troubleshoot any issues or problems that may arise.

– Cost: The video analytics software should be affordable and cost-effective, to fit your budget and provide a good return on investment. You should compare the pricing models and plans of different vendors, and consider the upfront and ongoing costs, such as the license fees, maintenance fees, and upgrade fees.

The cost per channel for video analytics software depends on the vendor, the features, and the plan that you choose. According to a report by IPVM, the average cost per channel for video analytics software ranges from $50 to $250, with some vendors offering discounts for bulk purchases or subscriptions. However, this is only an estimate, and you should contact the vendors directly to get a quote for your specific needs.

Video analytics is a powerful and promising technology that can help retailers improve their security and reduce their losses. However, video analytics also has some limitations and challenges that you should be aware of, such as:

– Accuracy: Video analytics is not 100% accurate, and may produce false positives or false negatives, depending on the quality of the video, the lighting conditions, the camera angles, the occlusions, and the variations in the appearance of the individuals or objects. You should always verify the results of the video analytics software with human judgment, and use other sources of evidence, such as receipts, inventory records, or witness statements, to confirm or refute the findings.

– Privacy: Video analytics may raise privacy and ethical concerns, as it involves collecting, processing, and storing personal data, such as faces, biometrics, or behaviors, of the individuals in the video. You should comply with the relevant laws and regulations, such as the General Data Protection Regulation (GDPR), and obtain the consent of the individuals before using video analytics on their data. You should also implement appropriate security measures, such as encryption, authentication, and access control, to protect the data from unauthorized access or misuse.

– Bias: Video analytics may exhibit bias or discrimination, as it may be influenced by the data, algorithms, or human factors that are used to train, test, or deploy the video analytics software. For example, the video analytics software may perform better or worse on certain groups of people, based on their age, gender, race, or ethnicity, due to the lack of diversity or representation in the data or the algorithms. You should be aware of the potential sources and impacts of bias in video analytics, and take steps to mitigate or eliminate them, such as using unbiased and representative data, testing and validating the software on different scenarios and populations, and monitoring and auditing the software for any signs of bias or discrimination.

I hope this article helps you understand how to use video analytics for retail security, and what are the best options and the limitations of this technology. If you have any questions or feedback, please feel free to contact me. Thank you for reading.