Our loyal four-legged friends’ emotions revealed in app

Imagine an app that tells you what your dog likes, which food tickles his taste buds and how he feels in different situations. In short, it’s an app that helps make his life almost perfect. The good news is such an app is on its way. The even better news is that you can play a part in its ongoing development!

Artificial intelligence is the buzzword of the moment. While some people associate AI with less pleasant matters, this technology offers many benefits, including for our four-legged companions. FaunaAI researchers Alina Hafner and Peter Gloor know this only too well. Together with other dedicated colleagues and animal lovers, they are working on an app that analyses animals’ emotions, including dogs.

Measuring emotions

“The concept of measuring animal emotions came to me when I tried to use my Happimeter to measure a horse’s emotions in real-time. It was a fruitless endeavour,” Gloor laughs. “The horse’s sheer size was a challenge, and its thick coat made it impractical to measure the animal’s emotional state through the wristband.”

So, what could serve as a reliable cue to gain a deeper insight into the animals? This line of thought immediately led Gloor to body language to decipher emotions. “Many people believe they can read their dog’s emotions. However, scientific research shows that only 50 per cent of people can accurately interpret a dog’s emotional state based on facial expressions,” he says. “The app we are developing aims to simplify the task for dog owners and help them understand the emotions and experiences of their loyal companion.”

Through its technology, the FaunaAI team aims to change the landscape of animal welfare by enabling comprehensive and accurate emotion analysis. “We use advanced technology that combines the latest developments in AI and machine learning to decipher and interpret the complex emotions of animals. Our algorithms analyse facial expressions, moods, body language and contextual cues,” explains Hafner.

More data

Although the first beta version of the app is already available, the FaunaAI team is actively seeking thousands of photos and videos to improve the training of the AI model. The more images available for analysis, the more accurate the app will be in analysing a dog’s emotional state.

“At the moment, there are about 600 to 700 photos and videos in the system, which is a sufficient starting point,” says Hafner. “However, we expect this number to increase significantly in the coming period, further refining the app’s ability to read dogs’ emotional states.”


The young scientist has already collected photos and videos in grooming salons, playgrounds and other places frequented by our canine companions. “I’ve noticed that people are sometimes wary of AI,” she notes. “They have expressed concerns about their dog’s privacy and whether the photos will be publicly available. Those who contribute can be assured that we treat all material with the necessary care”.

Photos and videos uploaded to the site are stored in a database accessible only to researchers. “This data will only be used to train and improve the AI model. The images themselves will not be accessible to the outside world,” Hafner clarifies.


Personal approach

For those interested in using the app, setting up a profile for their dog is simple. They can start immediately by uploading photos and videos for analysis. “This allows users to gradually gain a comprehensive understanding of their dog’s favourite places and the food that really excites him,” says Gloor.

But that’s not all, Hafner points out. “As well as providing an accurate and objective assessment of the dog’s emotions, the app can detect when the dog is unwell or in emotional distress. This allows the owner to take immediate action, ultimately contributing to the dog’s overall well-being.”

“Communication between dog and owner can also be improved,” says Gloor, “as the app encourages better mutual understanding and a deeper bond, leading to a stronger connection.”


Veterinarians, behavioural therapists, shelter staff and other dog-related professionals are invited to participate. “To accurately assess a dog’s happiness and well-being, we also need visual material of animals in less favourable circumstances. That is why we ask people who work professionally with dogs to contribute pictures,” explains Hafner.

“In the long term, we want to expand the range of emotions that the app can detect so that dog owners can understand how to give their loyal companion the most fulfilling life possible,” explains Gloor.