The University of the Philippines (UP) is making big strides in health and wellness with new AI-powered research projects that could revolutionize healthcare across the country. These groundbreaking ideas were recently showcased at AI Horizons PH 2024: A Conference on AI-Powered Research and Innovation, held on October 24-25, 2024, at the UP-BGC campus. From better diagnostics to personalized mental health therapy, these projects highlight AI’s potential to make healthcare faster, more accurate, and more accessible.
1. AI for Better Colonoscopy Screening
One major project is focused on improving colonoscopy screenings. Titled “Computer Vision and Deep Learning-Based Colonoscopy for Identifying Adenoma-Carcinoma Sequence Polyps,” this project uses AI to help doctors detect pre-cancerous polyps more accurately. The system works in real time, helping to reduce human error during colonoscopies, and is compatible with standard computers, making it affordable for more healthcare providers.
“With this AI tool, we can catch early signs of cancer more reliably,” explained Project Lead Dr. Anastacia Alvarez. This could lead to earlier treatments and better outcomes for patients.
2. Enhanced Dental Diagnostics
UP is also using AI to make dental care more accurate. In the project titled “Multi-class Semantic Segmentation of Tooth Pathologies and Anatomical Structures on Radiographs,” AI helps dentists identify dental issues more precisely using X-ray images. Built with a deep learning model called U-Net, this tool achieves a high level of accuracy, reducing the chances of missed diagnoses.
“This tool is designed to help dentists improve exam accuracy, supporting better oral health,” noted Project Lead Dr. Pros Naval.
3. Personalized Mental Health Therapy
Mental health care is another area where UP researchers are applying AI. The “Adaptive Mental Health Therapy” project uses a wearable device to monitor brain activity and deliver real-time, personalized therapy. By using AI-based neurofeedback, this technology creates custom mental health interventions that adapt to each person’s needs.
“Our goal is to make mental health therapy more accessible and tailored to individual needs,” said Project Lead Dr. Jaymar Soriano.
4. AI-Enhanced Radiology
The “DESERAY: Detection and Segmentation of Pneumothorax on Chest X-rays” project aims to help radiologists identify pneumothorax (collapsed lung) more accurately. This AI system detects the condition with high precision, reducing human error and speeding up the diagnosis process.
“DESERAY supports radiologists by improving accuracy and efficiency, which is critical for patient care,” explained Dr. Pros Naval.
5. Faster Heart Condition Diagnoses
UP’s “Two-View Left Ventricular Segmentation and Ejection Fraction Estimation” project uses AI to analyze heart health through echocardiograms more quickly and accurately. This tool can assess heart function, helping doctors diagnose conditions faster and with less expertise required.
“Our model can simplify cardiac diagnostics, making it easier to get accurate results quickly,” said Dr. Pros Naval.
6. Smarter Pathology with EDIAS
Another UP project, the “Enhanced Digital Image Analysis System (EDIAS),” assists pathologists by automatically interpreting images of tissue samples. This helps with diagnosing diseases like cervical cancer, especially in remote areas. EDIAS also supports mobile access, allowing doctors to send images for fast diagnosis from anywhere.
“This system speeds up and improves diagnoses, especially in areas lacking pathologists,” said Project Lead Dr. Alvin Caballes.
7. AI Tool for Prostate Cancer Detection
Lastly, UP’s “AI-Powered Explainable Decision Support Tool for Prostate Cancer Classification” project uses AI to help radiologists classify prostate cancer from MRI images. It also features an Explainability Dashboard, ensuring that medical professionals understand how the AI makes its decisions, encouraging trust and adoption in clinics.
“With this tool, we’re helping doctors make more informed, reliable decisions on prostate cancer cases,” said Project Lead Johanna Adevoso-Cañal.
