What is Image Recognition?
Image recognition is like teaching a computer to "see" and "understand" what's in a picture. It's a technology that allows a machine to look at an image and correctly identify things like objects, people, text, and more.
In Mentionlytics, image recognition helps you analyze the visual part of your mentions and extract knowledge on how your trackers appear (whether it's your brand or competitors, or even marketing leads) visually across social media, and enrich your social listening reports and dashboards with visual insights.
Think of it as giving the social listening tool eyes.
In this guide, you’ll find everything you need to know about Mentionlytics’ image recognition capabilities, Image Entity AI Analysis, and Image Search.
Image Entity AI Analysis
Mentionlytics’ AI algorithm scans all of your mentions with an image or a video thumbnail, analyzes them, and recognizes elements contained in them. We call each specific type of visual element an “entity”.
This could be your brand's logo (even if it's on a t-shirt or in the background).
It could be your product (like a specific type of soda bottle or shoe).
It can also identify the scene (a beach, a concert, a kitchen) or objects (a car, food, a computer).
Then, the system categorizes them based on the entities they contain.
The default image entities are: Logo, Object, People, Gesture, Landscape, Location, Event, Date, Text, and Country.
You can see the AI Image Recognition Entities if you:
Go to “Settings” on the left menu bar.
Click “Automations & Tags”.
Scroll to the bottom of the page.
However, please note that it’s for viewing only. You’re not able to alter them on your own.
Note: You can ask our customer support team to add any other entity you want for Mentionlytics to recognize within your mentions’ images, e.g., your product.
Important Clarification: Mentionlytics doesn’t scan the online space for images associated with the keywords you track. It collects mentions that contain your keywords and then analyzes the images contained within them. So, image recognition is conducted on the already collected mentions.
How to Add Image Entity AI Analysis to Your Reports
You can add an Image Analysis Entities slide in your custom reports from the report builder by following these steps:
Go to the report you want to include Image Analysis Entities in, or create a new report.
Click on “Add New Item” in the report builder.
Scroll down the options and select “Image Analysis Entities”.
The slide is now added to your report, and you can interact with it.
Add filters or styling options to that slide to tailor what you see.
How to Add Image Entity AI Analysis to Your Dashboards
You can add an Image Analysis Entities component to your custom dashboards, too. To do so, follow these steps:
Go to the custom dashboard you want to include the Image Analysis Entities component in, or create a new custom dashboard.
Click to “Enter Edit Mode” on the 3-dot icon
.
Click “+” where you want to add the component.
Scroll down the component options in the pop-up window.
Locate the “Image Analysis Entities” component.
Select it and customize it as you want with filters and styling options.
Click “Add Component & Close”.
The component is added to your dashboard, and you can interact with it.
👉 Learn more about the Custom Dashboards
Image Search
Once AI categorizes an image contained in a mention, it makes it searchable by that visual entity, even if the person who posted the picture didn't write any text about your brand or use a hashtag.
This means when you search for a specific entity, the platform filters your mentions to show only those visuals that contain it.
You can search through the images of your mentions for specific entities and find visuals that contain:
Your logo (or a competitor’s)
Objects (e.g., a smartphone, a coffee cup)
People (e.g., a chef, a customer)
Gestures (e.g., a thumbs up, a handshake)
Landscapes (e.g., a mountain, a city skyline)
Locations (e.g., Times Square, a local café)
Events (e.g., a concert, a press conference)
Dates (e.g., 06/11 or 01/01)
Texts (e.g., “New Collection” or “50% OFF”)
Countries (e.g., France, Japan)
To do this:
Go to “Filters”.
Select “Image Search”.
Choose one or more options from the drop-down list or type exactly what you want to find in the bar and click the little lens icon.
With this filtering method, you quickly narrow your mentions to only include the specific visual entities you are looking for.
In case you want to find visuals with a very specific entity in your mentions, e.g., the logo of your competitors, you can choose “Logo” and then type the name of your competitor in the little search bar that pops when you click the magnifying glass icon.
Advanced Image Filtering with Boolean Search
If you want to filter your mentions by multiple image entities, you can use Boolean Search.
Go to Boolean Search by clicking the 2-circle icon next to the search bar.
Locate the 2 new Boolean operators: entities.label: & entities.entity:
Use entities.label: to get results for mentions containing one of the pre-made entities (Logo, Gesture, Landscape, etc.) or for a custom entity you might have created with the help of your Account Manager. Remember to type the label exactly as it is written in your entity filters, with a capital first letter and the right spelling.
For example, if you wanted to see mentions of your logo on an object, in a specific country, this is what your Boolean query would look like.
Use entities.entity: to get results for mentions containing a pre-made or custom entity, but something else in particular. For example, if you wanted to see mentions of your logo on an airplane, this is what your Boolean query would look like.
Important: Remember to use Boolean queries only when you want to filter your data by multiple entities. Otherwise, it’s simple and handy to use the Image Search in filters.