This is happening today due to combination of visual data input, which reaches massive amounts, increasing graphic processing unit (GPU) power and improvements in deep learning methods. This blend gives power to analyse visuals and provide meaningful inferences like human brain processes them.
As Arçelik, we are working on computer vision applications particularly focusing on model compression in deep learning models running on edge devices as well as on cloud to scale up “Vision Intelligence Platform” in a most efficient way. We maintain our work to develop industry specific solutions for different market verticals; primarily retail, but also for schools, hospitals, workplaces, and production fields. For instance, we are able to count the incoming visitors who are visiting our stores but also analysing their demographic breakdown such as age group and gender to better understand in – store customer journey and whereabouts of visitors with heatmap analysis. Therefore, we are able to provide insights to relevant teams and executives by our retail analytics solution as a decision support tool. This makes our physical stores quantifiable and smart as if in the digital world to enhance customer experience.
Alongside with retail, we continue to work on application areas of face biometrics. By conducting face analysis of the person, as mentioned above one can determine gender detection, age group estimation, and emotional analysis of the person. It is worth to mention that we pay attention to privacy and processing face analysis anonymously in a compliant way. Furthermore, depending upon requirement of the use case scenario, we are able to implement other means of application areas by our face recognition engine such as verifying the claimed identity and authorizing access to designated zone(s) in order to create safer environment, standardized structure of doing business and measurable processes.
As described, Arçelik Vision Intelligence Platform focuses on retail analytics and face analysis solution sets however we continue to dig and explore for different implementations of computer vision field regardless of industry.