Search papers, labs, and topics across Lattice.
This paper introduces PelviNeXt, a hybrid network architecture that integrates a dense convolutional feature extractor, hierarchical channel-spatial attention, multi-scale fusion, and talking-heads multi-head self-attention to analyze pelvic imaging data from both ultrasound and X-ray modalities. The authors address the critical issue of data contamination in the PCOSGen dataset, providing a deduplicated version and establishing a new evaluation protocol, while achieving state-of-the-art performance on the PXR150 pelvic fracture dataset. The findings underscore the potential of a unified architecture to enhance diagnostic accuracy in under-researched areas of women's health, demonstrating its effectiveness across different imaging modalities without the need for task-specific adjustments.
A single architecture can achieve state-of-the-art results in pelvic imaging across multiple modalities, addressing critical data integrity issues in women's health research.
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.
