Organizations: Center for Computational & Data Sciences, Independent University, Bangladesh, Dhaka, Bangladesh · Department of Computer Science and Engineering, Independent University, Bangladesh, Dhaka, Bangladesh
Women's health remains substantially under-resourced in medical imaging research, with pelvic pathologies such as polycystic ovary syndrome (PCOS) and pelvic fracture both suffering from a scarcity of public, well-annotated benchmark data despite their clinical importance. We introduce PelviNeXt, a modality-agnostic hybrid architecture combining a dense convolutional feature extractor, hierarchical channel-spatial attention (H-CBAM), a multi-scale fusion module (MSFM), and talking-heads multi-head self-attention (TH-MHSA), applied without modification to both pelvic ultrasound and X-ray inputs. While benchmarking PelviNeXt on PCOSGen, the only gynaecologist-annotated public PCOS ultrasound dataset, we identified extensive exact and near-duplicate contamination within and across the dataset. We audit this contamination via perceptual hashing, publicly release a deduplicated version of the dataset, and establish the first integrity-audited evaluation protocol and baseline for PCOSGen under 5-fold cross-validation. On the only publicly available pelvic fracture X-ray dataset (PXR150), PelviNeXt exceeds previously reported state-of-the-art results across accuracy, recall, specificity, and AUROC. Ablation studies confirm that each architectural component contributes to performance on both tasks. Our results demonstrate that a single architecture, applied without task-specific modification, can serve as a reliable foundation for pelvic imaging across modalities in data-scarce, under-researched areas of women's health.
Figures & tables
Figure 1: Overview of the PelviNeXt architecture.
Dataset
Normal
Abnormal
Total
Original PCOSGen
1,319
3,349
4,668
After exact duplicate removal
948
2,523
3,471
After near-duplicate removal ( ≤14 )
63
162
225
Table 1: PCOSGen dataset composition before and after deduplication.
Figure 2: Example image pairs at increasing perceptual hash distance. (a) Distance 0: exact duplicate. (b) Distance 14: visually indistinguishable, used as the deduplication threshold. (c) Distance 16: visually distinct, confirming the threshold boundary.
Model
Acc.
Rec.
Spec.
F1
AUROC
ViT-B/16
89.33%
88.92%
79.36%
0.8337
0.8629
± 4.22
± 3.98
± 12.59
± 0.0802
± 0.0763
ResNet-101
88.44%
88.60%
80.90%
0.8427
0.8479
± 4.85
± 3.44
± 7.85
± 0.0402
± 0.0637
DenseNet-169
88.89%
87.14%
81.03%
0.8342
0.8418
± 3.64
± 11.90
± 5.81
± 0.0715
± 0.0664
Table 2: PCOS classification results on the deduplicated PCOSGen dataset. Mean over 5-fold CV; 95% CI shown in the row below each model’s results.
Model
Acc.
Rec.
Spec.
F1
AUROC
CLAHE [ 1 ]
80.67%
82.00%
80.00%
–
0.8140
Gamma [ 1 ]
80.67%
81.00%
81.00%
–
0.8160
Ensemble [ 2 ]
80.00%
82.00%
76.00%
–
0.7900
Patch Ensemble [ 2 ]
84.00%
87.00%
82.00%
–
0.8700
ViT-B/16
80.67%
84.00%
79.00%
0.8133
0.8210
± 5.62
± 3.67
± 6.50
± 0.0459
± 0.0535
Table 3: Fracture classification results on PXR150. Mean over 5-fold CV; 95% CI shown in the row below each of our models’ results.
Acc.(%)
F1
AUROC
PCOS
w/o H-CBAM
90.44% ± 1.48
0.8578 ± 0.0275
0.8758 ± 0.0196
w/o MSFM
88.89% ± 1.19
0.8509 ± 0.0190
0.8735 ± 0.0167
Vanilla MHSA
91.11% ± 2.75
0.8692 ± 0.0317
0.8851 ± 0.0238
PelviNeXt
92.00% ± 1.60
0.8890 ± 0.0085
0.9051 ± 0.0156
Fracture
Table 4: Ablation study. Mean ± 95% CI over 5-fold CV.
Figure 3: Grad-CAM visualizations from PelviNeXt. (a) Correctly predicted Abnormal PCOS ultrasound. (b) Correctly predicted Normal PCOS ultrasound. (c) Correctly predicted Fracture pelvic X-ray. (d) Correctly predicted Normal pelvic X-ray.
Pelvic fractures pose significant diagnostic challenges, particularly in cases where fracture signs are subtle or invisible on standard radiographs. To address this, we introduce PelFANet, a dual-stream attention network that fuses raw pelvic X-rays with segmented bone images to improve fracture classification. The network employs Fused Attention Blocks (FABlocks) to iteratively exchange and refine features from both inputs, capturing global context and localized anatomical detail. Trained in a two-stage pipeline with a segmentation-guided approach, PelFANet demonstrates superior performance over conventional methods. On the AMERI dataset, it achieves 88.68% accuracy and 0.9334 AUC on visible fractures, while generalizing effectively to invisible fracture cases with 82.29% accuracy and 0.8688 AUC, despite not being trained on them. These results highlight the clinical potential of anatomy-aware dual-input architectures for robust fracture detection, especially in scenarios with subtle radiographic presentations.
Center for Computational & Data Sciences, Independent University, Bangladesh · Department of Computer Science and Engineering, Independent University, Bangladesh · Advanced Medical Engineering Research Institute, University of Hyogo, Japan +1
Female pelvic diseases remain an under researched area characterized by often delayed diagnosis. While pelvic MRI offers superior soft-tissue contrast for diagnosis and image-guided procedures, real-time anomaly detection remains challenging due to physiological motion, tissue deformation, and instrument artifacts. Existing supervised approaches are impractical, as adverse events are rare, heterogeneous, and difficult to annotate. We present a Dinomaly-based unsupervised anomaly detection framework adapted for pelvic MRI that learns normative representations from healthy cases and flags deviations without requiring labels. Our approach leverages a frozen DINOv3 Vision Transformer encoder combined with a noisy MLP bottleneck and Linear Attention decoder to prevent identity mapping while maintaining computational efficiency. Anomalies are localized via per-token cosine distance between encoder and decoder representations, yielding spatial anomaly maps that provide immediate feedback at the scanner to support radiologist decision-making and adaptive protocol adjustment. Evaluated on a curated subset of the Uterine Myoma Dataset, the framework achieves a pixel-level AUROC of 88.06% and high specificity (95.45%) at frame level at 40.5 slices/s, meeting real-time clinical deployment requirements. The spatial anomaly maps and frame-level scores provide immediate, localized feedback at the scanner to support radiologist decision-making and adaptive protocol adjustment during active procedures.
Anika Knupfer, Maximilian Lindholz, Johanna Paula Müller +4
Institute of Information Processing, Leibniz University Hannover, Hannover, Germany · CAIMED, L3S, Hannover, Germany · Department of Radiology, Charité Universitätsmedizin Berlin, Berlin, Germany +2
Pelvic segmentation is one of the most important and fundamental research problems in precise and intelligent diagnosis and treatment, as well as surgical planning and navigation for pelvic fractures. By combining an improved geodesic active contour model with deep neural networks, we propose GUMP-Net, an interpretable model-data-driven intelligent algorithm for multi-class pelvic segmentation, in which three network modules are designed to constitute the overall segmentation framework together: the object detection module for automatic level set initialization, the edge detector module for learning an anatomy-aware edge detector function and the iteration module for deep level set evolution. Leveraging the advantages of level set representation and deep learning, GUMP-Net shows more accurate, robust and consistent segmentation performance, especially in small training data situation, compared to the state-of-the-art methods. Extensive experiments on pelvic datasets demonstrate the rationality and effectiveness of the proposed algorithm. Further experiments extended to ankle dataset indicate broader applications to other anatomies. The proposed algorithm not only provides an efficient segmentation method for complex fracture reduction, but also gives an interpretable geometric perspective for understanding deep learning segmentation.
Liheng Wang, Yinghui Zhang, Licheng Zhang +3
State Key Laboratory of Mathematical Sciences, Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing 100190, China; University of Chinese Academy of Sciences, Beijing 100190, China. · University of Chinese Academy of Sciences, Beijing 100190, China. · State Key Laboratory of Mathematical Sciences, Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing 100190, China. +4