CPL Anniversary Celebration Recognizes Research and Alumni Achievements
On October 1, 2022, the Computational Physiology Lab (CPL) celebrated the lab’s research achievements over the past two decades. The daylong event, held both in person and virtually, brought together over 40 current students, alumni, and collaborators. The celebration was full of stories, memories, lessons, and a distinct fondness for the lab and its founding director, Ioannis Pavlidis.
Physically present in the event were current Ph.D. students, doctoral and post-doctoral alumni who live in the Houston area, as well as alumni who traveled for the occasion from as far as North Carolina. Remote attendees included alumni and collaborators the world over, from Silicon Valley all the way to Maastricht, Holland, and reaching as far east as Amman, Jordan.
In the first part of the event, the lab’s alumni gave bittersweet accounts of their days in the lab, full of humor and nostalgia. The presentations proceeded in chronological order, painting a detailed historical picture of the lab’s evolution from an innovative hub of health informatics to a powerhouse in affective computing. In the second part of the event, Pavlidis presented an integrated account of the lab’s intellectual work and its impact on science and society. The event concluded in early evening with a dinner function in honor of the in-person participants.
Evolution of CPL
Pavlidis, who currently holds the Eckhard Pfeiffer Professorship in Computer Science, founded the lab in 2002 when he joined the University of Houston as an associate professor after working at Honeywell Laboratories for six years. His research interests intersect computing and human behavior, an area that back in the 2000s was so new that it did not have a name; today is known as affective computing. “Human behavior is very complex,” said Pavlidis. “It requires a 360 view and a convergence approach that brings together the fields of computing, psychology, and physiology.”
Initially, CPL focused on developing unobtrusive methods for sustained monitoring of physiological variables. Later, it applied these methods to study sympathetic hyperarousal (i.e., stress) and the problems it creates in human-machine and human-human interactions.
“We were the first to develop contactless vital sign measurement methods,” Pavlidis said. “There was nothing like that before us.” For instance, CPL developed a way to measure breathing and heart rates using thermal imaging cameras instead of chest belts and electrocardiograms. These revolutionary methods enabled the first naturalistic studies to measure and understand driving and workplace stress moment by moment.
As the technology advanced, so did CPL’s research methods. In recent years, the lab has performed some of the most cutting-edge affective computing studies, where multimodal data from personal devices and the environment are collected for days on end and subjected to artificial intelligence analysis. The results give unprecedented insights into daily micro-behaviors and phobias that have eluded explanation and even detection thus far.
Professional Impact of the Lab and Pavlidis on Members
The lab owes its success, in part, to the tireless efforts and unwavering dedication of the former members who have gone on to make remarkable contributions in both academia and industry. Pavlidis has graduated 21 Ph.D. students, six postdocs, and a dozen M.S. thesis students since joining UH’s Department of Computer Science. A few alumni shared their personal stories and memories of CPL, highlighting the professional impact that both the lab and Pavlidis had on them.
Working in the CPL allowed Ashik Khatri to pursue his passion for building software tools alongside learning and conducting research. After Khatri’s positive mentorship by Pavlidis in the UH Computer Science Research Experience for Undergraduates summer program, Khatri knew he wanted to pursue a graduate degree with the same mentor. “One thing that truly caught my eye is that Professor Pavlidis was able to highlight the strengths of his students and provide appropriate guidance where needed,” said Khatri. During his time with CPL, he developed several mobile, web, and desktop apps to support multiple research projects. “The long hours spent debugging, designing, testing, and enhancing software applications were some of the most challenging times that helped me become a better software engineer today,” he said. Khatri is currently managing director at Glowderma.
CPL’s unique, fascinating research, and its practicality is what attracted Dinesh Majeti, now a senior software engineer at Teradata Corp., to the lab as a Ph.D. student. “It is difficult to find an application to your [research] work as a Ph.D. student,” Majeti said. “But the projects in CPL are very experimental and have real-world applicability. I loved the work, so I joined the lab.” A fun and noteworthy project that Majeti recalls is running experiments at the gym where test subjects walked different speeds and inclines on a treadmill with an iPhone in hand. The goal of the project was to estimate the surface incline using the accelerometer data collected from the iPhone to improve mobile walking apps. “I was surprised to see how much data the iPhone was collecting,” he said.
An Infrared Imaging course taught by Pavlidis sparked Dvijesh Shastri’s interest in joining CPL. “His teaching material exposed us to his innovative research projects,” said Shastri. “I also enjoyed his cool and approachable nature, which for me was indicative of a good research mentor.” Shastri primarily worked on a deception detection project that aimed to detect stress levels using thermal imaging of individuals interrogated for mock crimes. This project was the highlight of his time with CPL because it was featured on CNN’s The Situation Room with Wolf Blitzer and The Discovery Channel. Shastri is currently a professor of Computer Science at the University of Houston–Downtown.
When Nanfei Sun joined CPL as a Ph.D. student, he helped build the Automatic Thermal Monitoring System. ATHEMOS allows for touchless measurements of human physiological signals such as blood and breathing rates. “Technological innovations of the time allowed us to capture subtle thermal signal changes with higher resolutions that were not available before,” said Sun. “Our team developed novel algorithms and built bio-heat models simulating the heat transfer process, to extract vital signs from thermal signals.” One of Sun’s career highlights since graduating was working at IBM Thomas J. Watson Research Center to develop a smart surveillance system used at Chicago O’Hare International Airport and the 2008 Summer Olympics in Beijing. Sun now works as assistant professor of Management Information Systems at the University of Houston–Clear Lake.
As an undergraduate, Yan Zhou became interested in computer vision research. “CPL drew my attention because it had an interesting lie detection research project that used computer vision technology,” said Zhou. Currently a principal engineer at FutureDial, Inc., Zhou attributes her career success to the experience she gained with CPL under Pavlidis’ mentorship. “[He] showed me excellent writing and communications skills…and those skills made me go a long way in the workplace,” she said. “Over my 11-year career path, I have steadily expanded my job responsibilities, grown from a research scientist to senior staff engineer, and now a principal engineer.”
At the Forefront of Health Informatics and Affective Computing Research
CPL’s impact stands tall and grows bigger with time. Contactless physiological measurements, first conceived by Pavlidis in the early 2000s, have found applications in health informatics, where Clair Labs and other high-tech start-ups have been busy developing homebound, contactless physiological monitoring that promises to revolutionize healthcare. During the COVID-19 pandemic, Boston Dynamics released the Dr. Spot canine robot, equipped with thermal and visual cameras for contact-free physiological measurements in clinical settings. CPL’s studies on driving distractions have received unprecedented coverage and are considered the affective computing standard. One can Google the phrases “sixth sense in driving” or “driving distractions pavlidis” to get a glimpse of the broad impact of CPL’s work in this area.
Over the years, CPL researchers consistently published in iconic journals and conferences, including Nature, Nature Human Behavior, Science Advances, and ACM CHI; in the latter, CPL has published nearly every year for two decades – an impossible record for a venue with ~20% acceptance rates.
The anniversary event was a fitting tribute to the laboratory’s achievements over the past 20 years. It highlighted the CPLers’ foundational role in the new disciplines of affective computing and health informatics. Pavlidis expressed his gratitude to current students, alumni, and collaborators – all of whom have contributed to the success of the lab – and looks forward to continuing to be at the forefront of convergence research for years to come.
Learn more about the groundbreaking research of the Computational Physiology Lab at https://cpl.uh.edu.
- Ioannis Pavlidis, Matthew Pariyothorn, and Ny’Ree Butler, Department of Computer Science