|
Gilching, Germany, April 14, 2026 — Allied Vision today announced that its exo183CGE CMOS camera served as the core imaging hardware in a peer-reviewed study on upright radiotherapy patient positioning, a medically critical discipline where millimeter-scale errors can carry clinical consequences. The study, conducted by researchers at Japan’s National Institutes for Quantum Science and Technology (QST) and published in the Journal of Applied Clinical Medical Physics, represents the first quantitative assessment of setup accuracy across multiple immobilization configurations for both abdominal and head and neck (HN) cancer treatments in the upright posture.
Upright radiotherapy, in which patients sit or stand on a rotating platform while fixed horizontal beams are delivered, offers significant advantages over conventional supine (lying down) treatment. These include reduced respiratory motion, smaller facility footprints, and dramatically lower infrastructure costs, particularly for charged particle therapies such as carbon ion radiotherapy. Yet the technique has been limited in clinical adoption, in part because the setup accuracy of upright patient positioning had never been rigorously quantified across multiple body regions and immobilization methods.
To solve the measurement challenge, QST researchers deployed three Allied Vision exo183CGE 20-megapixel GigE Vision color cameras — one front-facing and one on each lateral side — in a trinocular stereo configuration surrounding a replicated clinical upright radiotherapy chair. Paired with 25 mm focal-length optical lenses and configured at f/4 for a depth of field exceeding 50 cm, the imaging system acquired synchronized 5496 × 3672-pixel frames of 26 surface markers placed on subjects’ bodies and immobilization masks.
Image acquisition across the three-camera array was synchronized to a median time lag of just 5 milliseconds. This high level of temporal precision is essential for accurate triangulation of 3D marker positions. The cameras were controlled via the open-source Harvester image acquisition library, and the resulting point clouds were processed through intrinsic and extrinsic calibration pipelines.
The result: a validated 3D surface imaging system with mean absolute position errors not exceeding 0.32 mm per axis and 3D position errors no greater than 0.45 mm, delivering submillimeter accuracy across the full measurement volume, verified against a precision grid plate.
WHAT THE DATA REVEALED
Enabled by the imaging system, the QST study delivered clinically actionable findings across six distinct patient setup configurations, tested on 15 volunteers over repeated sessions:
- Inter-fractional setup accuracy: Use of thermoplastic masks in abdominal setups reduced mean 3D setup error from 6.6 ± 3.3 mm to 3.9 ± 1.7 mm. The optimal head and neck configuration achieved 2.9 ± 1.7 mm, the lowest error among all HN setups tested.
- Intra-fractional displacement: The best-performing setups held intra-fractional motion to 1.9 ± 1.1 mm (abdominal) and 1.8 ± 1.5 mm (head and neck) over 20-minute treatment sessions. These results are statistically equivalent to published benchmarks from supine radiotherapy.
- Beam delivery time: The optimal abdominal setup maintained tolerable displacement thresholds for a mean of 17.8 ± 2.7 minutes out of a 20-minute window — a 87% duty cycle that demonstrates real-world clinical viability.
SIGNIFICANCE IN CLINICAL RESEARCH
The exo183CGE’s high pixel count, geometric consistency, and reliable synchronization across multi-camera configurations made it uniquely suited for this application. Optical surface imaging in radiotherapy demands cameras that can maintain calibration integrity over extended experimental sessions, resolve sub-5mm markers at distances of 50+ cm, and deliver images with sufficient bit depth and contrast for automated centroid detection — requirements the exo183CGE met without compromise.
This study establishes a new methodological benchmark: by using triangulation from multiple high-resolution camera views rather than 2D optical guidance tools employed in prior work, the QST team achieved full 6-degrees-of-freedom motion tracking across both body and mask surfaces simultaneously. That capability is only possible with cameras that deliver consistent, high-fidelity imaging frame after frame.
ABOUT ALLIED VISION
Allied Vision stands for customer-centric 2D Machine Vision solutions, from individual components to fully integrated complete solutions. With deep application knowledge from various technology areas and industries, Allied Vision develops tailored vision solutions that enable customers to make informed decisions based on precise information. The product portfolio ranges from line scan and area scan cameras to lenses and frame grabbers, and from image processing libraries to IP cores – all from a single source. As a trusted partner with a global network of experts and local support, Allied Vision stands by its customers and partners, making collaboration as easy and reliable as possible.
Allied Vision is part of the Machine Vision division of the TKH Group. In January 2026, Allied Vision, Chromasens, Mikrotron, NET, and SVS-Vistek were consolidated under the Allied Vision brand.
|
