Researchers from Korea’s one of the best science and technology universities, KAIST’s Department of Bio and Brain Engineering, have developed a new artificial intelligence-powered light field camera that can read 3D facial expressions.
The highly capable camera uses a technique that uses infrared light to read facial expressions. Professors Ki-Hun Jeong and Doheon Lee led the research team which developed this artificial intelligence-enabled technology.
The newly developed light-field camera comes with micro-lens arrays in front of the image sensor, allowing it to capture the spatial and directional information of light in a single shot, making it tiny enough to fit into a smartphone.
Read More: TCS, Wipro, Infosys to carry out Research at JNTU-Kakinada
The research named ‘Machine-Learned Light-field Camera that Reads Facial Expression from High-Contrast and Illumination Invariant 3D Facial Images’ was published in Advanced Intelligent Systems.
Prof Ki-Hun Jeong said, “The sub-miniature light-field camera developed by the research team has the potential to become the new platform to quantitatively analyze the facial expressions and emotions of humans.”
He further added that the technology could be used in various fields like mobile healthcare, field diagnosis, social cognition, and human-machine interactions. According to the researchers, facial expressions captured through 3D images had more than 80% accuracy, making it a reliable technology.
The camera uses a near-IR-range vertical-cavity surface-emitting laser (VCSEL) to stabilize the precision of 3D picture reconstruction, which was previously dependent on ambient light. Researchers say that the light-field camera eliminates 54 percent of image reconstruction errors when an external light source is placed on a face at 0-, 30-, and 60-degree angles.
Researchers were also able to reduce optical crosstalk while boosting picture contrast by 2.1 times. This was made possible by introducing a light-absorbing layer for visible and near-IR wavelengths between the micro-lens arrays.