P. Rizwan Basha*
Chidananda H.
Nandish K
Pamidi Naheer
Pregnancy is a sensitive phase that demands safe physical activity, continuous monitoring, and personalized healthcare guidance to ensure the well-being of both mother and fetus. Although medical experts recommend regular exercise during pregnancy, improper posture and lack of supervision can lead to health risks. Most existing pregnancy and fitness applications provide static exercise videos or general guidelines without real-time posture correction or intelligent monitoring. With recent advancements in smart computing and artificial intelligence, particularly in computer vision–based pose estimation, it has become possible to analyze human movements accurately in real time. This paper presents a Pregnancy Guide and Exercise Monitoring System, an AI-powered web-based application that utilizes MediaPipe Pose to perform real-time posture detection and repetition counting during pregnancy-safe exercises while ensuring user privacy through client-side processing. In addition to exercise monitoring, the system offers trimester-specific pregnancy guidance, nutrition recommendations, health tracking, automated report generation, and notification services. A role-based access mechanism enables secure interaction among patients, doctors, and administrators, supporting remote monitoring and informed healthcare management through an intelligent and scalable digital platform.
PROBLEM STATEMENT
Pregnant women often lack access to personalized guidance and real-time supervision while performing physical exercises, which can lead to incorrect posture and potential health risks. Existing pregnancy and fitness applications mainly provide static information and general exercise videos without intelligent monitoring, posture correction, or pregnancy-specific safety feedback. Additionally, these systems offer limited support for continuous health tracking and remote monitoring by healthcare professionals. Therefore, there is a need for an intelligent, smart computing–based solution that ensures safe exercise monitoring, personalized pregnancy guidance, and secure interaction between patients and healthcare providers.
OBJECTIVES
- To design and develop an AI-powered pregnancy care system for safe exercise monitoring.
- To implement real-time posture detection and repetition counting using computer vision techniques.
- To provide instant feedback and safety alerts during pregnancy-safe exercises.
- To offer trimester-specific pregnancy guidance and nutrition recommendations.
- To enable health tracking, automated report generation, and notifications.
- To support secure role-based access for patients, doctors, and administrators.
LITERATURE REVIEWS
|
Authors |
Year |
Title / Focus |
Merits |
Remarks |
|
Lugaresi et al. |
2019 |
MediaPipe: A Framework for Perception Pipelines |
Enables real-time pose estimation with high accuracy |
Not specific to pregnancy care |
|
Bazarevsky et al. |
2020 |
BlazePose: Real-Time Body Pose Tracking |
Lightweight and efficient pose tracking |
Requires domain-specific customization |
|
Thompson et al. |
2022 |
Computer Vision–Based Exercise Monitoring |
Improves exercise posture and feedback |
General fitness, not pregnancy-focused |
|
Patel and Mehta |
2022 |
AI-Based Remote Health Monitoring Systems |
Supports personalized and remote healthcare |
Limited real-time activity monitoring |
|
ACOG |
2022 |
Exercise During Pregnancy Guidelines |
Provides medically approved exercise practices |
No AI or monitoring support |
|
Kumar et al. |
2023 |
AI in Maternal Healthcare |
Highlights AI potential in pregnancy care |
Lacks practical implementation |
|
Alharbi et al. |
2024 |
Attention-Based Video Summarization Models |
Enhances activity understanding |
Not designed for healthcare use |
|
Ge et al. |
2024 |
Smart Computing in Healthcare Systems |
Demonstrates smart system integration |
Does not include posture monitoring |
METHODOLOGY
The proposed system follows an AI-driven methodology to ensure safe and personalized pregnancy care. Live video input from the user’s camera is processed locally using MediaPipe Pose to perform real-time posture detection and repetition counting during pregnancy-safe exercises, ensuring privacy by avoiding video transmission to the server. The frontend, developed using React, provides user interaction and exercise guidance, while the backend, implemented using Django REST Framework, manages authentication, data processing, and role-based access. User data, exercise records, and health information are securely stored in a relational database. Additional modules provide pregnancy guidance, nutrition recommendations, health tracking, reports, and notifications, enabling effective remote monitoring and smart healthcare support.
DESIGN
- Context diagram
2.Use case diagram
3.Sequence diagram
RESULTS AND DISCUSSION
Starting Page
The starting page introduces the AI-Powered Pregnancy Care Application and guides users to begin the system through login or registration and provides three login options: Patient Login, Doctor Login, and Admin Login, allowing users to access the system based on their roles.
Patient Login
Allows patients to securely log in and access exercise monitoring, nutrition guidance, and health reports.
Doctor Login
Enables doctors to log in and view patient health data and exercise reports in read-only mode.
Admin Login
Allows administrators to access system management, user control, and analytics features.
Fig1: starting
Fig 1.1 : login window
Patient Login
The patient login provides secure access for registered users to the AI-Powered Pregnancy Care Application. After successful authentication, patients can view and update their personal and pregnancy details, perform AI-monitored pregnancy-safe exercises, access trimester-specific nutrition guidance, receive notifications and reminders, and view weekly health reports. This login ensures personalized care while maintaining data privacy and security.
Fig 1.2: patient Home page
Fig 1.3: Nutritional Guide
Fig 1.4: Exercise Library
Fig 1.5: monitoring sys
Fig 1.6: activity data upload
Fig 1.7: weekly health report
Fig 1.8: My Profile
Fig 1.9:Notification
Fig 2.0:Custom Reminders
DOCTOR LOGIN
The doctor login allows authorized healthcare professionals to securely access the system. After logging in, doctors can view patient profiles, monitor exercise performance, review health reports, and analyze pregnancy-related data in read-only mode. This feature supports remote patient monitoring while ensuring data security and integrity.
Fig 2.1:Doctor Dashboard
ADMIN LOGIN
The admin login provides secure access for system administrators to manage the overall platform. After logging in, administrators can manage users, monitor system activities, view analytics, and ensure smooth and secure operation of the application.
Fig 2.2: Admin Dashboard
CONCLUSION
This paper presented a Pregnancy Guide and Exercise Monitoring System that leverages smart computing and artificial intelligence to support safe and personalized pregnancy care. By integrating computer vision–based pose estimation, the system enables real-time exercise monitoring, posture correction, and repetition counting while preserving user privacy through client-side processing. In addition, the application provides trimester-specific guidance, nutrition recommendations, health tracking, and notification services with secure role-based access for patients, doctors, and administrators. The proposed solution demonstrates the effective use of emerging AI technologies in maternal healthcare and offers a scalable approach for enhancing exercise safety, remote monitoring, and overall pregnancy well-being.
REFERENCES
- G. Lugaresi et al., “MediaPipe: A Framework for Building Perception Pipelines,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, Seattle, WA, USA, pp. 1–9, June 2019.
- V. Bazarevsky et al., “BlazePose: On-device Real-time Body Pose Tracking,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, Nashville, TN, USA, pp. 1–9, June 2020.
- TensorFlow Team, “TensorFlow.js: Machine Learning for the Web,” TensorFlow Documentation, Google, 2023.
- Django Software Foundation, “Django: The Web Framework for Perfectionists with Deadlines,” Official Django Documentation, 2024.
- T. McGinnis and J. Kind, “Django REST Framework: Powerful and Flexible APIs,” Django REST Framework Documentation, 2024.
- World Health Organization (WHO), “WHO Guidelines on Physical Activity and Sedentary Behaviour,” Geneva, Switzerland, 2020.
- American College of Obstetricians and Gynecologists (ACOG), “Physical Activity and Exercise During Pregnancy and the Postpartum Period,” ACOG Committee Opinion, no. 804, 2022.
- R. B. Thompson et al., “Computer Vision–Based Exercise Monitoring for Healthcare Applications,” IEEE Access, vol. 10, pp. 112345–112356, Jan. 2022.
- S. Patel and A. Mehta, “AI-Based Remote Health Monitoring Systems: A Survey,” Journal of Medical Systems, vol. 46, no. 3, pp. 1–15, Mar. 2022.
10.5281/zenodo.19755139