PATIENT POSITION DETECTION

Accurate patient positioning is a critical requirement in diagnostic radiography, as improper positioning and involuntary patient movement can result in image degradation, repeated exposures and increased radiation dose. This project presents the design and development of a low-cost, marker-free Patient Position Detection system intended to assist radiographers in monitoring patient stability and verifying the correct radiographic view before X-ray exposure. The proposed system employs a camera-based computer vision approach using Python, OpenCV and MediaPipe Pose and Hands to detect and track anatomical landmarks across five body regions: head, chest, hips, leg and hand. The system monitors patient movement within the collimator field of interest by comparing real-time landmark coordinates against a predefined reference position. Motion severity is estimated using positional displacement, changes in apparent body size and movement velocity. Additionally, view classification is performed to distinguish between Anteroposterior (AP), Posteroanterior (PA), lateral and oblique views using landmark geometry and estimated rotation angles. An integrated traffic-light feedback mechanism provides visual alerts, indicating stable positioning, slight movement or significant movement and view mismatches. An HLK-LD2401 millimetre-wave radar sensor is incorporated as an additional sensing component to measure front-to-back displacement, complementing the camera-based monitoring approach. The prototype also supports motion event logging through CSV files and allows for future integration with IP cameras. Live testing was conducted using a laptop camera, focusing on hand positioning. The system successfully identified the expected AP and PA views in all five recorded test cases, with corresponding traffic-light indications. The results demonstrate the feasibility of real-time, camera-assisted patient position monitoring. Further validation across all five body regions, integration of radar measurements into the primary monitoring interface and clinical testing are required before practical deployment. Keywords: Patient Position Detection, Computer Vision, Radiography, MediaPipe, OpenCV, Pose Estimation, Motion Detection, View Classification, mmWave Radar.

Authors

Institutions

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-19
DOI
https://doi.org/10.5281/zenodo.24374423
Primary Topic
Radiation Dose and Imaging
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

PATIENT POSITION DETECTION

B.M.S. College of Engineering, Sriharshaa Padmakumar
Zenodo (CERN European Organization for Nuclear Research)
Radiation Dose and Imaging
preprint

PATIENT POSITION DETECTION

B.M.S. College of Engineering, Sriharshaa Padmakumar
preprint en

Abstract

Accurate patient positioning is a critical requirement in diagnostic radiography, as improper positioning and involuntary patient movement can result in image degradation, repeated exposures and increased radiation dose. This project presents the design and development of a low-cost, marker-free Patient Position Detection system intended to assist radiographers in monitoring patient stability and verifying the correct radiographic view before X-ray exposure. The proposed system employs a camera-based computer vision approach using Python, OpenCV and MediaPipe Pose and Hands to detect and track anatomical landmarks across five body regions: head, chest, hips, leg and hand. The system monitors patient movement within the collimator field of interest by comparing real-time landmark coordinates against a predefined reference position. Motion severity is estimated using positional displacement, changes in apparent body size and movement velocity. Additionally, view classification is performed to distinguish between Anteroposterior (AP), Posteroanterior (PA), lateral and oblique views using landmark geometry and estimated rotation angles. An integrated traffic-light feedback mechanism provides visual alerts, indicating stable positioning, slight movement or significant movement and view mismatches. An HLK-LD2401 millimetre-wave radar sensor is incorporated as an additional sensing component to measure front-to-back displacement, complementing the camera-based monitoring approach. The prototype also supports motion event logging through CSV files and allows for future integration with IP cameras. Live testing was conducted using a laptop camera, focusing on hand positioning. The system successfully identified the expected AP and PA views in all five recorded test cases, with corresponding traffic-light indications. The results demonstrate the feasibility of real-time, camera-assisted patient position monitoring. Further validation across all five body regions, integration of radar measurements into the primary monitoring interface and clinical testing are required before practical deployment. Keywords: Patient Position Detection, Computer Vision, Radiography, MediaPipe, OpenCV, Pose Estimation, Motion Detection, View Classification, mmWave Radar.

Zenodo (CERN European Organization for Nuclear Research)
B.M.S. College of Engineering (IN)
Radiation Dose and Imaging
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon