Cam QA Tool helps cut USD $59,000 costs in face recognition

3DiVi has introduced the Cam QA Tool, an automated solution to assess camera positioning and configuration in video analytics projects that depend on face recognition.

Face recognition performance in video analytics systems can be hindered by issues such as improper camera placement, suboptimal lighting, and environmental factors, including dust and glare. These problems may not be identified until after a system is deployed, potentially resulting in reduced accuracy, failed identifications, and project overruns that increase costs.

3DiVi's Cam QA Tool has been developed to address these challenges by providing automated analysis of both live and recorded camera feeds. The tool evaluates 19 specific parameters that directly influence face recognition performance. It then generates a comprehensive report containing detailed findings and practical recommendations aimed at improving camera setups during both initial installation and post-deployment phases.

The company states that the Cam QA Tool offers key benefits for both system owners and operators. It provides an objective method for validating face recognition performance at the acceptance stage and tracking any decline in quality as the system is used.

System integrators can use the tool to reduce the risk of client dissatisfaction related to inadequate face recognition. Additional applications include selecting suitable equipment for various analytics tasks, monitoring and evaluating installation teams, optimising workflow across projects, and enhancing the technical capabilities of operational teams.

The process involves uploading either a live camera feed via RTSP stream or a 30-minute video sample. The Cam QA Tool then reviews footage to analyse between 30 and 50 unique faces for optimal assessment. Following the analysis, users receive a PDF report indicating the status of the 19 measured performance parameters and offering potential improvements where necessary.

In a real-world application, a video management system integrator utilised the Cam QA Tool to assess its camera inventory. The evaluation revealed that 4K cameras did not enhance face recognition accuracy over lower-resolution alternatives. This finding allowed the team to switch to more affordable 2K and Full HD cameras, resulting in cost savings of USD $59,000, equivalent to 17% of the project's total budget, without a negative impact on performance.

Another case study provided by 3DiVi highlighted the use of Cam QA by a public safety department. A year after camera installation, a post-deployment review identified factors such as dust accumulation, lighting deficiencies, and minor alignment issues that had dropped face recognition accuracy to 47%. Upon implementing the recommendations, the department increased accuracy to 96%, improving identification outcomes in ongoing cases.