A team of researchers at Edith Cowan University (ECU) in Western Australia has made a breakthrough in artificial intelligence (AI), developing a system that better understands human emotions, paving the way for more empathetic and responsive machines.
Unlike traditional AI models that rely on analyzing single facial images, the ECU team introduced a method that interprets emotions by examining sets of related facial expressions. This multi-frame approach mimics how humans assess emotions over time, offering a richer, more nuanced understanding.
“Just like we don’t judge how someone feels from one glance, our method uses multiple expressions to make more informed predictions,” said Sharjeel Tahir, ECU PhD student and lead author of the study.
The technique enhances AI accuracy, even under variable lighting conditions or when faces are viewed from different angles. By training AI with diverse visual cues grouped in structured sets, the system not only becomes more precise but also remains computationally efficient, noted co-author and fellow PhD student Nima Mirnateghi.
Although not yet integrated into physical robots, the technology is expected to impact emotionally intelligent systems in mental health care, customer service, and education globally. The researchers are now focused on developing artificial empathy, allowing AI to respond appropriately to human emotional states and make decisions that are easier for users to understand.
“Our work lays the foundation for machines that can truly understand and respond to people’s feelings,” said Dr. Syed Afaq Shah, senior lecturer at ECU and lead researcher on the project.
The findings were detailed in a release published Wednesday on the Science Media Exchange.
Written By Rodney Mbua
