Wrist bone segmentation in X-ray images using CT-based simulations

Alexia Karantana; Xin Chen; Youssef ElTantawy

REVIEW 4 major objections 5 minor 16 references

Wrist bone segmentation in X-ray images using CT-based simulations

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Using synthetic X-rays made from CT scans, a deep learning model segments eight wrist bones plus radius and ulna in real radiographs across view angles from -70 to 70 degrees.

desk verdict Simulation-trained multi-view wrist segmentation is a solid proof-of-concept, but the paper's quantitative evidence is self-referential and the real-X-ray transfer claim remains unvalidated. read the letter →

arxiv 2507.07131 v1 pith:KEFJMPNB submitted 2025-07-08 eess.IV cs.CVq-bio.TO

Youssef ElTantawy , Alexia Karantana , Xin Chen This is my paper

classification eess.IVcs.CVq-bio.TO

keywords wristbonesegmentationX-rayimageCT-basedsimulationsynthetictrainingdatasim-to-realtransfercarpalbonesdeeplearningmulti-viewradiographs

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to show that a deep learning model can learn to segment the eight carpal bones plus the distal radius and ulna in ordinary wrist X-rays without any hand-labeled X-rays for training. Instead, the authors generate thousands of synthetic radiographs by projecting CT volumes at angles from -70 degrees to 70 degrees, using the CT bone labels as ground truth masks. On simulated test images the model reaches mean Dice scores of about 0.92 near the front-back view and 0.80 at 70 degrees; on 121 real radiographs the results are judged qualitatively, not by numbers. If the sim-to-real transfer holds, this removes the main bottleneck of expert annotation for multi-view wrist segmentation.

What carries the argument

The load-bearing mechanism is an orthographic X-ray simulation that models each pixel as the exponential of the negative summed attenuation along a ray through the CT volume, treating the CT values as absorption coefficients and ignoring scatter and energy dependence. Ground-truth masks come from 3D multi-label bone segmentations of the CT scans, obtained for one neutral pose and propagated to all poses by a statistical shape-and-kinematic wrist model; projected labels are resolved to the closest bone when overlap occurs. A tissue-reduction step clipped low intensities to boost bone contrast, with thresholds chosen by trial and error, and augmentation by rotation, translation, zoom, and horizontal flip widens the simulated distribution. Training then uses a self-configuring encoder-decoder network with combined Dice and cross-entropy loss.

What would settle it

Have a radiologist manually label the eight carpal bones and two forearm bones on the 121 real radiographs (or register the CT volumes to them to get ground truth) and compute Dice scores; if the real-image scores fall far below the simulated range of 0.80 to 0.92, the central claim of sim-to-real transfer fails.

Watch

Extended reading notes

Core claim

The central claim is that segmentation trained on CT-based simulated X-rays generalizes to real X-rays for ten wrist bones across a wide range of view angles. The paper reports that a single model, trained on 9240 synthetic images from 88 CT volumes, produces average Dice scores of 0.92 at 0 degrees and 0.80 at 70 degrees on a simulated test set, with per-bone values ranging roughly from 0.67 to 0.95. On real radiographs, the same model visually produces acceptable outlines and preserves anatomical overlap relations even in lateral views, though it fails when only part of the hand is visible. The authors state that no existing method automatically segments all eight carpal bones plus the two forearm bones in arbitrary views.

Load-bearing premise

The whole approach depends on the assumption that the simplified simulated X-rays---parallel rays, no scatter, no energy dependence, and tissue intensities clipped by trial-and-error thresholds---look enough like real clinical radiographs that a model trained on them transfers to real images.

Editorial extensions

If this is right

Wrist segmentation training no longer needs manually annotated X-rays; CT-derived masks can supply arbitrarily many training pairs.
The same model covers view angles from -70 degrees to 70 degrees, including lateral views where carpal bones overlap heavily, something prior methods restricted to front-back views.
The released model and simulation code let others generate synthetic wrist X-rays for downstream tasks such as fracture detection, bone age assessment, and arthritis evaluation.
Performance degrades with view angle, with average Dice falling from 0.92 near the front-back view to 0.80 at 70 degrees, so high-angle lateral views remain the weak end.
Left and right hands are handled through horizontal flip augmentation, removing a manual mirroring step.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If the sim-to-real transfer is as good as the qualitative real-image results suggest, the same recipe could be applied to other joints or to the metacarpal bones, wherever CT volumes with bone labels exist.
A quantitative real-image benchmark, produced by expert annotation or by 3D-to-2D registration of the CT volumes to the radiographs, would be the natural next test and would tell how much of the simulated Dice survives the gap.
Adding perspective projection, scatter, and beam-hardening effects to the simulator, rather than handling them with augmentation and clipping, may specifically improve the high-angle and overlap-heavy cases where Dice drops below 0.8.
The dependence of the tissue-reduction thresholds on trial and error suggests a sensitivity analysis: reporting how segmentation varies as those thresholds move would show how robust the pipeline is.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The paper does something genuinely useful: it trains a nnU-Net to segment all eight carpal bones plus the distal radius and ulna in wrist X-rays across view angles from -70 to 70 degrees, using 9240 simulated radiographs generated from CT volumes. That is a real advance over prior work, which either used manual labels on AP views only (Kang et al.) or excluded the carpals entirely (Lee et al.). The writing is clear, the motivation is sound, and the authors are honest about the limits of their evaluation.

The core weakness is that the quantitative evidence for the central claim comes entirely from simulated test images produced by the same pipeline used to generate training data. Table 1 reports Dice and ASD on 912 simulated images from held-out CT scans, but those images share the same projection model, the same tissue-reduction thresholds, and the same augmentation scheme as the training set. So the scores mostly show self-consistency, not transfer to real radiographs. The real-image evaluation, Section 3.3, is explicitly qualitative: 121 X-rays, no ground truth, selected examples, and no error rate or failure analysis beyond noting that partial hands fail. That is a missing-evidence problem, not an internal inconsistency, but it leaves the load-bearing claim—that simulation-trained segmentation transfers to clinical X-rays—unvalidated.

There are also smaller concerns. The simulation itself is deliberately simplified (orthographic, scatter-free, monoenergetic), and the tissue-reduction step uses thresholds determined by trial and error; these choices could plausibly distort bone contrast and edges in ways that matter for real images. The paper does not compare against any existing method, although it is fair to say that no prior method handles all ten bones across view angles. And despite the abstract promising the trained model and simulation code "freely available... link will be provided upon acceptance," nothing is public at the moment, which limits independent verification.

On the positive side, the paper is not overclaiming in its conclusion: it explicitly says the method was evaluated quantitatively on simulated images and qualitatively on real images, and it defers quantitative real-image validation to future work. The related work is cited appropriately, including the authors' own prior work on CT segmentation and 3D statistical models, which is directly relevant and not self-citation for its own sake.

Who gets value: researchers working on sim-to-real transfer in medical imaging, or on wrist or hand segmentation, will find this a useful proof-of-concept and a clear statement of the evaluation gap. It deserves a serious referee: the question is real, the method is reasonable, and the paper is honest. But as it stands, the central claim is not yet established. My recommendation is to send it to peer review with the expectation of major revision, requiring quantitative validation on real X-rays with reference labels, a baseline comparison, and public release of code and data.

Referee Report

4 major / 5 minor

Summary. The paper proposes to train a deep-learning segmentation model (nnU-Net) on simulated wrist X-ray images generated from CT volumes, with the goal of automatically segmenting eight carpal bones plus the distal radius and ulna in real plain radiographs across view angles from -70 to 70 degrees. The simulation uses orthographic projection with simplified attenuation physics, and the corresponding masks are obtained by projecting the CT bone segmentations. The authors report Dice scores of 0.80-0.92 on a simulated test set, and present qualitative segmentation results on 121 real wrist radiographs without ground truth.

Significance. If the transfer from simulated to real X-rays were quantitatively established, the approach would offer a valuable solution to the label-scarcity problem for multi-view wrist segmentation, potentially enabling downstream applications such as fracture detection and bone age assessment. The paper's strengths include the use of a publicly available training pipeline (code and model), a relatively large synthetic dataset covering multiple angles, and the inclusion of all eight carpal bones. The central claim, however, rests on a self-referential evaluation: the quantitative metrics are obtained only on simulated images generated by the same pipeline used to create the training data, while the real-image evaluation is explicitly qualitative. The significance of the reported Dice scores therefore depends entirely on the unvalidated assumption that the simulation distribution is close enough to real clinical radiographs.

major comments (4)

[Section 3.3 and Section 4] The central claim that the model segments real X-ray images is not quantitatively supported. The abstract states that the method was evaluated using both simulated and real images, but Section 3.3 presents only qualitative examples for real radiographs, and Section 4 explicitly defers quantitative real-image evaluation to future work. Since the entire quantitative evidence (Table 1) comes from simulated test images produced by the same pipeline as the training data, the reported Dice/ASD scores demonstrate self-consistency of the synthetic generation procedure rather than generalization to real X-rays. Please provide a quantitative evaluation on real radiographs with expert-annotated ground truth for at least a subset of the 121 images, or alternatively use a public wrist X-ray segmentation benchmark, in order to substantiate the abstract's claim.
[Section 2.2, Eq. (1)] The realism of the simulation is load-bearing for the paper's transfer claim, yet the simulation is heavily simplified and includes ad hoc adjustments. Specifically, the orthographic, scatter-free, monoenergetic model in Eq. (1) ignores key physical effects, and the tissue-reduction step uses thresholds 'determined by trial and error' (setting intensities below the 20th percentile to the 10th percentile). These choices can plausibly distort bone contrast and edges relative to real radiographs. To gauge the sim-to-real gap, the paper should include at least a quantitative comparison of intensity statistics or a small calibration experiment on real images with labels; without such evidence, the simulated Dice scores alone do not establish that the model will perform on real X-rays.
[Section 3.1] The statement 'No existing method could segment wrist radiographs in a variety of view angles, hence only our method was evaluated' is not a valid justification for omitting comparison or quantitative evaluation. Even if no prior method covers all the claimed view angles, existing wrist-segmentation methods (e.g., Kang et al. [4]) could be compared on the AP subset, or the authors could report a reader study on the real radiographs. At a minimum, the real-image segmentation results should be quantified with metrics that do not require dense ground truth, such as boundary agreement between two expert annotators or overlap-based scores on a small manually labeled subset.
[Section 2.1] The training and test labels are not fully independent of the method being evaluated. The multi-label masks for all CT scans are obtained using the interactive segmentation tool [6] plus the shape and kinematic models from [7]; the same masks are used to generate both the training labels and the simulated test-set ground truth. Any systematic error in this label-propagation pipeline will inflate the reported Dice scores because the model is trained and tested on the same label definitions. The paper should report validation of the propagated masks against manual annotations on non-neutral poses, or at least quantify inter-observer variability of the segmentation masks.

minor comments (5)

[Table 1] The bone name 'Treapezium' appears to be a typo for 'Trapezium' in Table 1 and in Section 3.2; please correct it.
[Abstract] The abstract states 'Dice scores ranging from 0.80 to 0.92' without specifying that this range corresponds to average Dice per view angle; the per-bone Dice at 70° can be substantially lower (e.g., Capitate 0.67 in Table 1). Please clarify the range in the abstract to avoid overstating the results.
[Section 2.3] The term 'nnUnet' should be consistently written as 'nnU-Net' to match the reference [10] and the standard notation.
[Section 3.1] The sentence 'the results of both negative and positive of the same angle are combined' is unclear; please specify whether this refers to averaging left and right wrists or to combining positive and negative rotation angles, and how the averaging is performed.
[Abstract and Section 5] The statement that the code and model 'link will be provided upon acceptance' is non-standard for a journal submission; please provide a stable repository or detailed instruction for obtaining the materials in the final version.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the simulated-image evaluation is a standard held-out-subject test within the synthetic distribution, and the real-image results are explicitly qualitative.

full rationale

The paper's derivation chain is not circular. CT volumes are rendered into synthetic X-rays via Eq. (1), and the corresponding masks are projected from 3D CT segmentations (Section 2.2). The model is trained on 9240 simulated images from 18 subjects and tested on 912 simulated images from 4 held-out subjects (Section 3.1). This is a held-out-subject evaluation within the simulated distribution, not a prediction forced by construction: the test CT scans and their rendered images are not seen during training, and performance degrades with view angle, so the reported Dice scores carry real information about the model's behavior on the simulator. The real-image evaluation (Section 3.3) is explicitly qualitative, with no ground truth, and the conclusion acknowledges that quantitative real-X-ray evaluation is future work: 'We will include more simulated conditions to improve the model performance and produce a dataset to quantitatively evaluate real X-ray images.' That is a limitation of external validity (the sim-to-real gap is not measured), not an internal circularity. The use of the authors' prior work [6,7] to generate CT bone labels is a reuse of published methods, and the paper does not invoke a self-citation to define its target in terms of its inputs or to forbid alternatives. Therefore no circular step is exhibited.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The central claim rests on the realism of the simulation, the accuracy of propagated CT labels, and the representativeness of the augmentation; the paper states the first as an assumption, determines the tissue-reduction threshold by trial and error, and provides no external validation of the label-propagation ground truth.

free parameters (4)

Tissue reduction percentile clamp = 20th percentile mapped to 10th percentile
Chosen by trial and error in Section 2.2 to mimic vendor post-processing; directly affects simulated bone/tissue contrast and therefore segmentation performance.
Artifact correction percentile cap = 99th percentile
Ad hoc upper clip in Section 2.2 to suppress metal-ring artifacts; affects the intensity distribution of simulated images.
Augmentation ranges = rotation +/-40 degrees, translation up to 20%, zoom +/-20%, horizontal flip
Hand-selected in Section 2.3; central to the model's view-angle robustness and to the simulated test set.
View angle sweep = -70 to +70 degrees in 10-degree steps
Design choice in Section 2.2 that defines the pose coverage; not fitted to data but determines the claimed flexibility.

assumptions (6)

domain assumption Parallel-beam exponential attenuation model with negligible scatter and monoenergetic absorption (Eq. 1)
Stated in Section 2.2; real X-ray systems use cone-beam geometry, polyenergetic spectra, and scatter, so this approximation may create a domain gap.
domain assumption CT values can be treated as attenuation coefficients and min-max normalized without distorting anatomy
Assumed in Section 2.2; the mapping from CT Hounsfield units to absorption is linear only approximately.
domain assumption The interactive segmentation masks [6] and their propagation via the statistical kinematic model [7] are accurate enough to serve as ground truth for all 107 CT volumes
Masks are generated for the neutral pose and propagated; no quantitative validation of mask accuracy is reported.
domain assumption The five wrist positions plus augmentation cover the clinically relevant distribution of real wrist radiographs
Relied on in Sections 2.1 and 2.3; if real clinic poses fall outside this range, the model may fail.
domain assumption Tissue reduction assumes each simulated image contains pure tissue regions
Stated in Section 2.2; when false, the percentile-based contrast enhancement distorts intensities.
standard math nnU-Net with Dice plus cross-entropy loss is a suitable segmentation backbone
Cited as state-of-the-art [10]; standard usage, not an ad hoc assumption.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Wrist bone segmentation in X-ray images using CT-based simulations." pith.science (2026). https://pith.science/paper/KEFJMPNB

@misc{pith2026250707131,
  author       = {Pith},
  title        = {Pith review of: Wrist bone segmentation in X-ray images using CT-based simulations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KEFJMPNB}},
  note         = {Machine review of arXiv:2507.07131}
}

read the original abstract

Plain X-ray is one of the most common image modalities for clinical diagnosis (e.g. bone fracture, pneumonia, cancer screening, etc.). X-ray image segmentation is an essential step for many computer-aided diagnostic systems, yet it remains challenging. Deep-learning-based methods have achieved superior performance in medical image segmentation tasks but often require a large amount of high-quality annotated data for model training. Providing such an annotated dataset is not only time-consuming but also requires a high level of expertise. This is particularly challenging in wrist bone segmentation in X-rays, due to the interposition of multiple small carpal bones in the image. To overcome the data annotation issue, this work utilizes a large number of simulated X-ray images generated from Computed Tomography (CT) volumes with their corresponding 10 bone labels to train a deep learning-based model for wrist bone segmentation in real X-ray images. The proposed method was evaluated using both simulated images and real images. The method achieved Dice scores ranging from 0.80 to 0.92 for the simulated dataset generated from different view angles. Qualitative analysis of the segmentation results of the real X-ray images also demonstrated the superior performance of the trained model. The trained model and X-ray simulation code are freely available for research purposes: the link will be provided upon acceptance.

Figures

Figures reproduced from arXiv: 2507.07131 by the authors.

**Figure 2.** which cover different view angles, and different fieldof-views for both left and right hands. It can be seen that the model produced acceptable segmentation results and correct anatomical relationships between the overlapped bones, even at very challenging view angles. The method failed when only part of the hand was visible (Fig.2 red box), as this situation was not simulated in the training process. It can be app… view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

16 extracted references · 15 canonical work pages

[4]

8 carpal bones and 2 distal forearm bones)

CONCLUSION A wrist X -ray image segmentation method has been presented, which can achieve automatic segmentation of ten bones (i.e. 8 carpal bones and 2 distal forearm bones). The method utilized thousands of simulated X -ray images and their corresponding masks generated from CT scans that cover different view angles, field -of-views, gender, and age gro...

work page
[6]

CONFLICTS OF INTEREST This work was supported by the Medical Research Council, U.K., under Grant 87997

work page
[7]

Automatic Segmentation of Ulna and Radius in Forearm Radiographs

Gou, X., Rao, Y., Feng, X., Yun, Z., and Yang, W. Automatic Segmentation of Ulna and Radius in Forearm Radiographs. Comput. Math Methods Med. 6490161, 2019

work page 2019
[1]

However, imaging the body using X-rays is complicated and sometimes ambiguous to be interpreted due to that 3D anatomic structures overlap each other in the 2D image

INTRODUCTION Plain X-ray image (also known as 2D radiograph) is utilized in a wide variety of medical applications, such as bone fracture detection, pneumonia diagnosis, and cancer screening, due to its relatively low cost, quick image acquisition time, and capability to see -through the body . However, imaging the body using X-rays is complicated and som...

work page
[2]

METHOD 2.1. CT Dataset and Real X-ray Images The CT dataset contains 22 subjects (10 female and 12 male, median age 51, age range 25 - 72 years) recruited from the hand clinic at Salford Royal Hospital, Greater Manchester, UK. Each subject was imaged at five different wrist positions: neutral, and four extreme positions in radial-ulnar and flexion extensi...

work page
[3]

EVALUATION 3.1. Experimental Design Among the 22 unique subjects (107 CT scans) , 88 CT volumes from 18 subjects were used for model training, and 19 CT scans from the remaining 4 subjects were reserved for testing. In total, 9240 (i.e. 88×15×7) simulated images were used for model training. The test set was generated using the same process as the trainin...

work page
[5]

Ethical approval was obtained for the use of this data

COMPLIANCE WITH ETHICAL STANDARDS This research study was conducted retrospectively using human subject data made available through local medical sources. Ethical approval was obtained for the use of this data. The subjects used in this study had consented to be included in this research. All data was anonymized, and the participants’ information cannot b...

work page
[8]

Multimed

Faisal, A., Khalil, A., Chai, H.Y ., and Lai K.W., X-ray carpal bone segmentation and area measurement. Multimed . Tools Appl. 81, 37321–37332, 2022

work page 2022

Show all 16 references

[9]

, Automatic inference and measurement of 3D carpal bone kinematics from single view fluoroscopic sequences

Chen X ., Graham J ., Hutchinson C ., and Muir L. , Automatic inference and measurement of 3D carpal bone kinematics from single view fluoroscopic sequences. IEEE Trans . Med. Imaging, 32(2):317-28, 2013

2013
[10]

J Pers Med., 12(5):776, 2022

Kang B.K., Han Y., Oh J., Lim J., Ryu J., Yoon M.S., Lee J., and Ryu S., Automatic Segmentation for Favourable Delineation of Ten Wrist Bones on Wrist Radiographs Using Convolutional Neural Network. J Pers Med., 12(5):776, 2022

2022
[11]

, Osteoporosis Prediction from Hand and Wrist X -rays using Image Segmentation and Self-Supervised Learning

Lee, H., Hwang, U., Yu, S., Lee, C., & Yoon, K. , Osteoporosis Prediction from Hand and Wrist X -rays using Image Segmentation and Self-Supervised Learning. ArXiv, abs/2311.06834, 2023

2023 arXiv
[12]

E., Integrated framework for simultaneous segmentation and registration of carpal bones, the 18th IEEE International Conference on Image Processing (ICIP), pp

Chen X., Graham J., and Hutchinson C. E., Integrated framework for simultaneous segmentation and registration of carpal bones, the 18th IEEE International Conference on Image Processing (ICIP), pp. 433-436, Belgium, 2011

2011
[13]

, Automatic generation of statistical pose and shape models for articulated joints

Chen X., Graham J ., Hutchinson C ., and Muir L. , Automatic generation of statistical pose and shape models for articulated joints. IEEE Trans Med Imaging. 33(2):372-83, 2014

2014
[14]

Phys Med Biol

Chen X ., Varley M .R., Shark L .K., Shentall G .S., and Kirby M.C., A computationally efficient method for automatic registration of orthogonal x-ray images with volumetric CT data. Phys Med Biol. 21;53(4):967-83, 2008

2008
[15]

In Maier A ., Steidl S ., Christlein V ., et al., editors

Berger M ., Yang Q ., and Maier A. In Maier A ., Steidl S ., Christlein V ., et al., editors. Medical Imaging Systems: An Introductory Guide, Springer; 2018

2018
[16]

F., Kohl, S

Isensee, F., Jaeger, P. F., Kohl, S. A., Petersen, J., and Maier - Hein, K. H., nnU-Net: a self-configuring method for deep learning- based biomedical image segmentation. Nature Methods, 18(2), 203- 211, 2021

2021

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.

[1] [4]

8 carpal bones and 2 distal forearm bones)

CONCLUSION A wrist X -ray image segmentation method has been presented, which can achieve automatic segmentation of ten bones (i.e. 8 carpal bones and 2 distal forearm bones). The method utilized thousands of simulated X -ray images and their corresponding masks generated from CT scans that cover different view angles, field -of-views, gender, and age gro...

work page

[2] [6]

CONFLICTS OF INTEREST This work was supported by the Medical Research Council, U.K., under Grant 87997

work page

[3] [7]

Automatic Segmentation of Ulna and Radius in Forearm Radiographs

Gou, X., Rao, Y., Feng, X., Yun, Z., and Yang, W. Automatic Segmentation of Ulna and Radius in Forearm Radiographs. Comput. Math Methods Med. 6490161, 2019

work page 2019

[4] [1]

However, imaging the body using X-rays is complicated and sometimes ambiguous to be interpreted due to that 3D anatomic structures overlap each other in the 2D image

INTRODUCTION Plain X-ray image (also known as 2D radiograph) is utilized in a wide variety of medical applications, such as bone fracture detection, pneumonia diagnosis, and cancer screening, due to its relatively low cost, quick image acquisition time, and capability to see -through the body . However, imaging the body using X-rays is complicated and som...

work page

[5] [2]

METHOD 2.1. CT Dataset and Real X-ray Images The CT dataset contains 22 subjects (10 female and 12 male, median age 51, age range 25 - 72 years) recruited from the hand clinic at Salford Royal Hospital, Greater Manchester, UK. Each subject was imaged at five different wrist positions: neutral, and four extreme positions in radial-ulnar and flexion extensi...

work page

[6] [3]

EVALUATION 3.1. Experimental Design Among the 22 unique subjects (107 CT scans) , 88 CT volumes from 18 subjects were used for model training, and 19 CT scans from the remaining 4 subjects were reserved for testing. In total, 9240 (i.e. 88×15×7) simulated images were used for model training. The test set was generated using the same process as the trainin...

work page

[7] [5]

Ethical approval was obtained for the use of this data

COMPLIANCE WITH ETHICAL STANDARDS This research study was conducted retrospectively using human subject data made available through local medical sources. Ethical approval was obtained for the use of this data. The subjects used in this study had consented to be included in this research. All data was anonymized, and the participants’ information cannot b...

work page

[8] [8]

Multimed

Faisal, A., Khalil, A., Chai, H.Y ., and Lai K.W., X-ray carpal bone segmentation and area measurement. Multimed . Tools Appl. 81, 37321–37332, 2022

work page 2022