Real-Time Menstrual Discomfort Detection Using AI-Powered Facial Recognition in Structured Environments

Main Article Content

Venkatesh Koreddi, Vinaya Sree Bai Kshatriya, Darapaneni Bhavishya, Chandaka Gowtami, Kolapalli Divya Sree, Konidena Anitha

Abstract

Women’s involvement, focus, and ability to function in structured environments like conferences, offices, colleges, and schools are all impacted by menstrual discomfort. However, the menstruation issue, which results in a shortage of accommodations in a timely manner, is not addressed by current appearance and access control systems. We propose a unique AI- managed menstruation detection method that employs closed-circuit (CC) cameras for facial recognition to solve this problem. Our method uses deep learning to analyze physiological indicators and facial expressions, including changes in skin texture, eye fatigue, puffiness, and discomfort associated with menstruation, to determine the individual’s condition. This guarantees a non-intrusive, privacy-preserving approach to providing immediate support in professional and academic contexts. Menstrual Presence Facial Dataset (MAFD-2024), a custom dataset comprising facial images taken one day prior to and five days following menstruation onset, is introduced to facilitate this research. It serves as one of the primary indicators of menstrual pain and has been carefully annotated. To obtain an accurate health identification model, we enhance it into a hybrid CNN-LSTM model 1 (HCL-MD), achieving a 94.1% improvement in both temporal (symptom progression) and spatial (facial characteristics) accuracy. Additionally, we integrate this technique into an automated framework for granting permission, enabling real-time adaptation for affected individuals. This technology can be incorporated into telemedicine systems in the healthcare industry, allowing physicians to remotely evaluate sufferers’ levels of discomfort and provide immediate remedies. Additionally, this technology can be used by smart wearables or surveillance systems in structured settings like conferences, schools, or public transportation to provide real-time support, including suggesting wellness treatments or break periods. By addressing menstruation discomfort in real time without the need for verbal communication, this AI-driven method not only raises awareness but also promotes a more compassionate and inclusive atmosphere. This work is at the forefront of integrating facial recognition with menstrual health tracking.

Article Details

Section
Articles