The Digital Economy Development Institute (DEDI LLC, part of the CDC Group), resident of the Skolkovo Foundation’s IT Cluster, has released a new version of the SmartLook Engine image recognition platform using artificial intelligence. The company's specialists have significantly increased the speed and accuracy of recognizing goods on store shelves using photos directly in a mobile device. Now, in less than a second, you can achieve a guaranteed recognition accuracy of 98%, even on conventional and inexpensive smartphones.

SmartLook Engine is an innovative platform for machine learning models, created on the basis of its own methods and recognition engine.

According to Khuseyn Az-zari, DEDI’s General Director, today it is the only image recognition system in Russia that analyzes the image in a smartphone or tablet even without an Internet connection.

Khuseyn Az-zari emphasized, “Russian and foreign counterparts are software that is installed on a mobile device, sends photos and videos to servers or cloud services and receives recognition results with a significant time delay.”

The SmartLook Engine platform can be used to develop image recognition systems in various industries. The solution is already being used for retail, allowing you to check the availability and quality of the placement of goods, analyze the shelf share for each category of goods, and timely update the price tags on the shelves. The system also analyzes POS materials and specialized stands in retail outlets. The SmartLook Retail image recognition system based on the SmartLook Engine platform greatly simplifies and speeds up the work of merchandisers, auditors and sales representatives.

In industry, the recognition platform can be used to develop quality control systems on an assembly line or monitor equipment malfunctions. In healthcare, the platform can be used for the primary diagnosis of certain diseases by photography and for the analysis of medical images. In agriculture, the system will help detect pests, detect plant diseases, and control crop maturation.

Khuseyn Az-zari said, “The scope of our engine is extensive: it can be embedded in mobile applications and video surveillance systems to solve various problems in many sectors of the economy. For example, an application based on the SmartLook Engine identifies employees to access the premises. Overseas retailers also use a solution to monitor user sentiment. Leading domestic retailers are also moving in this direction: an assessment of the emotional state of buyers is carried out when tasting products."

Recall that in 2019, the Mars began implementing an automatic product recognition system using artificial intelligence on the SmartLook Engine platform in the Central Eurasia, Belarus and Turkey region.

Due to its unique technological advantages, the SmartLook Engine solution allowed Mars to obtain reliable marketing indicators at any point of sale in a matter of seconds, regardless of the quality of the mobile connection. As a result of the implementation, the operating time of merchandisers in retail outlets was reduced by about 25%, which ensured a significant increase in the coverage of the territory and significantly increased the presence of Mars in the countries of the region.

More than 70% of the data that affects the assessment of the quality of work with points of sale and sales volumes is filled in automatically based on the analysis of photos in a mobile device. With the SmartLook Engine image recognition platform, Mars is getting 5x more marketing data than before. Now the assessment of the representation of goods is carried out in the context of each commodity item (SKU) both for own goods and for goods of local and international competitors, without increasing the cost of obtaining such information. The use of the SmartLook Engine has allowed Mars to remain in the lead in 2020 and increase its share of sales despite the restrictions associated with the pandemic.

Currently, more than 800 mobile employees of Mars are working with the SmartLook Retail in 7 countries of the region (Kazakhstan, Belarus, Kyrgyzstan, Uzbekistan, Armenia, Georgia, Azerbaijan). In the coming year, it is planned to introduce it in other countries of the region - in Moldova, Mongolia, Tajikistan, Turkmenistan, and Turkey.