Goods on the shelf recognition
ArtLabs team developed the solution for goods recognition on the product shelves. Our solution helps sales representatives and merchandisers to automatically compare the planned and actual display on the shelf at the point of sale. It is enough for a merchandiser to take a photo of the shelf, and he will be able to inspect the calculated share of the brand, which positions are missing and other analytics directly on the phone.
We used latest advancements in Computer Vision to develop this Software, including State-Of-The-Art Object Detection and Image Classification algorithms.
Deep Learning Algorithms
After our solution was implemented, there were a few simple steps to set it up. We developed an API which accepted an image and outputted all the recognized products and their corresponding metrics
When developing our solution we experimented with various State-Of-The-Art algorithms for Object Detection and Image Classification, including EfficientNet and EfficientDet algorithms. In the end we selected the based model based on the inference time and the corresponding metrics.
Features Of The API
Our API can be easily integrated with the merchandisers’ app. Here are a few features our customer was able to access after the API configuration: Number of products for each SKU, Missing products, Product placement, Product Prices, Layout Quality and many more
Once the API our customer could add new products via API directly. There was no need to contact us to do the additional work.