Ad display optimized for customers visiting the store

店舗/施設毎の来店者数カウントや、顧客の性別・年齢等の属性を推定、さらにデジタルサイネージなどで掲示する広告コンテンツの視聴時間を計測します。これによりどの広告がどれだけ見られたか（視認性の計測）や、どの顧客がどの広告に興味を示したか（属性毎の効果）など、広告効果の可視化を可能にします。これらの情報を基にターゲット顧客層毎に最適化した広告を実現できます。

また、サイネージでは顧客属性(年齢/性別)に応じたコンテンツの切り替えやターゲット層からの広告視認率の見える化によるインプレッション課金にも活用が可能です。

Layout optimized for target customers

Count the number of people in a store / facility, measure the length of stay, and estimate the attributes of visitors. This makes it possible to measure which route to follow from entering the store (tracking of leads) and where many people gather for each attribute (people flow analysis). Based on this information, it is possible to optimize the layout according to the purchasing behavior of the target customer.

infrastructure development

Place cameras in various places to prepare an environment that can acquire on-site information. In addition, edge AI devices are installed to perform AI analysis of images from cameras with higher accuracy, and analysis is performed at a location close to the site.

Video analysis on the edge side not only improves accuracy, but also reduces costs such as data communication charges and provides security measures for personal information.

Data analysis and consideration of measures

Link information such as the number of people / attributes acquired at a store / facility with customer/sales information accumulated in core systems such as POS data. We will analyze the acquired data and consider the measures to be taken.

Implementation and improvement of moves

Based on the analysis results, we will reset the target, improve the advertisement display and store layout, etc., and continuously improve while repeating the effect measurement each time.

camera

Existing surveillance cameras can also be used. The installation location, number of units, and angle of view are determined according to the data to be acquired and the AI software to be installed.

Video analysis software/hardware

AI software that matches what you want to achieve (purpose and accuracy) is not developed in-house from scratch, and there are many products that have been learned. In order to operate these, an edge computer for AI analysis called an edge AI device is required. It is preferable to have a built-in GPU specialized for video analysis.

Choosing the right tools

There are so many manufacturers and lineups of AI software and edge AI devices that you can choose according to your application. We will select products that have a track record of introduction in Japan and overseas, and products that match the purpose and accuracy. The important thing is to first define the requirements for what you want to achieve, list candidate tools, and examine their feasibility.

PoC (demonstration experiment) and identification of issues

After the on-site staff understands how to use the system, trial operation begins. Make adjustments based on issues you notice during operation.

Actual operation

We will start operating the production system at the actual store in a state where the software version upgrade after operation and the maintenance system for the equipment are in place. Based on the information obtained by video AI, you can improve the service.