
Upgrade your analysis with Landmark Recognition
Not only logo or label recognition, but landmark detection too. KPI6 is continuing to empower the listening of online conversations thanks to the use of the best performing and most exclusive AI features. How does Landmark recognition work in “Black Lives Matter” protests’ social media contents? Let’s take a look.
Would you like to know how a Digital Consumer Intelligence approach can help your Business? Book a Demo to get a free hands-on demonstration of our platform.
When we take a photo, are we really conscious of what we photograph?
Our faces, our characteristics, our outfit. A sunset, a lake, a beautiful beach. Everything we photograph remains in one click … and can be analyzed!
Well yes, nothing goes unnoticed to artificial intelligence!
As previously seen in past articles, the new Artificial intelligence features manage to label the visual contents of an image, manage to define with precise labels if there are faces, products, but also brands and monuments or artistic places. Thanks to this machine learning and AI tools, it is possible to understand where a photo was taken, to recognize squares or cities, without resorting to the analysis of geographical coordinates.
A Landmark detection feature that KPI6 cannot absolutely miss and that has integrated with its Digital Consumer Intelligence technologies.
Google Vision Landmark Detection
This feature, as well as logo recognition, uses Google Vision technologies. Google Vision API is a particular type of API that allows developers to analyze the content of images, using machine learning models. Thanks to this technology, we can have contextualized information on a given image and be able to quickly classify images in various categories and sub-categories.
Google Vision API allows you to identify the content of an image with great precision, to provide the name of the landmark, a confidence score and a bounding box in the image for the landmark and to give coordinate for the detected entity.
KPI6 has decided to connect Google Vision technologies to its platform to make more performing the classification of online conversations and to provide its customers with enriched insights, from textual to a visual level.
Landmark detection in “Black Lives Matter” protests
After the killing of George Floyd in Minneapolis, USA, hundreds of thousands of people took to the streets in the last two weeks in various cities around the world to participate in anti-racism demonstrations inspired by the ‘Black Lives Matter’ movement.
Black Lives Matter (BLM) is an international human rights movement that campaigns against violence and systemic racism towards black people. BLM regularly holds protests speaking out against police brutality and police killings of black people.
The demonstrations began in the USA and continued in many countries around the world, including Italy, where there were marches, flash mobs and peaceful rallies over the past weekends.
By analyzing the Twitter conversations based on the hashtag #blacklivesmatter, KPI6 managed to understand where the conversations were mainly concentrated, not taking advantage of the geolocation of each device, but only by the landmark detection use.
Thanks to this Artificial Intelligence technology, it was possible to “read” the images, analyze them, and understand what were the main monuments, streets and squares photographed during the demonstrations.
The first step was to listen to all the online conversations about these protests. We created a search query on our platform that included references to the main hashtag used by protesters (#blacklivesmatter), including keywords about the protest, movement, demonstration, community, peaceful revolution and many more…
The online conversations generated by listening have been empowered by the AI feature in the classification phase, in order to have a perfect overview of the landmarks recognized in users’ photos.
Thanks to locations Insight, it was possible to show the geographic distribution of the research. The map shows the posts that were tagged in a location. As we can see, the focus of the protests was in the USA, on both coasts. There have also been many demonstrations in the United Kingdom, particularly in London.

And until now, no additional news. The peaceful (sometimes not) demonstrations were widely reported in all the newspapers. But, is it possible to understand, analyzing the social contents, where did the demonstrations take place, not taking geolocation into account? Absolutely yes, thanks to the use of Landmark Recognition.

This particular feature managed to identify specific places by analyzing the content of the photos. That’s how we found out that most of the protests took place in Washington DC, near the White House and President’s Park. Many other East Coast locations have been recognized, such as Rhode Island State House or Lincoln Memorial.

In the West Coast’s protests, the technology managed to recognize some very important and historical points of Los Angeles and San Francisco: The Hollywood Walk of Fame and the Golden Gate Bridge.


Landmark recognition has become an integrative feature in conversation monitoring actions. It is possible to use this technology to monitor the participation in some important events, to monitor the photographs of shop windows for a fashion industry company, to associate a particular food with a particular place. Multiple matching that can be useful in the creative and strategic phase.
Would you like to know how a Digital Consumer Intelligence approach can help your Business?
Book a Demo to get a free hands-on demonstration of our platform.