Enrich your digital customer data with camera-based physical behavior analysis to gain deeper insights.
Deriving insights from the physical world is important to accurately and deeply understand your customers for better marketing activities, while data-driven marketing mainly focuses only on their digital behaviors such as web access logs and online purchase histories.
To give an example, when a person spent more than a minute in front of a specific shelf at a retail store, it could be a strong indicator of their interests. Bringing such real-world information to Treasure Data makes a big difference to your day-to-day data-driven business activities by enriching conventional digital datasets.
SCORER is a lightweight, easy-to-use camera-based human behavior analysis solution running on both edge and cloud. The solution allows you to easily capture and analyze the physical world (e.g., customer behaviors in retail stores) by using camera devices. Eventually, this Box demonstrates a way to connect the solution to Treasure Data, and you will be able to obtain deeper insights into your customers by integrating their digital and physical footprints.
It should be noted that there is a trade-off between edge-side and cloud-based video analysis in terms of efficiency, accuracy, and richness of the outputs:
Choose an appropriate option (or, even hybridize the implementation) based on your requirements.
SCORER Edge: Simple, Efficient Analysis Running on the Edge
SCORER Edge is a Raspberry Pi-based lightweight camera device that enables you to easily capture and analyze video stream on the edge.
Edge-side video analysis is effective to significantly reduce network traffic and quickly extract actionable insights from incoming data because the devices do not need to send raw video data to the cloud and wait for computationally expensive operations. By contrast, this option may pose some difficulties in terms of accuracy and richness of the outputs due to the limited computing resources allocated to the edge devices.
SCORER Cloud: Rich, Advanced Cloud-Based Video Analysis Platform
SCORER Cloud is a cloud-based video analytics platform that allows us to analyze real-world human behaviors from video streams captured by camera devices.
Unlike edge-side video analysis, cloud-based solution enables you to efficiently leverage advanced techniques at ease. For instance, one of the most sophisticated techniques named SenseVideo estimates demographics of people captured in a video; it is computationally too expensive to run on SCORER Edge, and Edge is usually suitable for simpler tasks such as counting up the number of pedestrians.
- Customer Enrichment
- People-Flow/Counting Analysis
- In-Store Shopper's Behavior, Shelf Monitoring