A Deep-Learning Approach for Vocal Fold Pose Estimation in Videoendoscopy DOI Creative Commons
Francesca Pia Villani, Maria Chiara Fiorentino, L. Fédérici

et al.

Deleted Journal, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 12, 2025

Abstract Accurate vocal fold (VF) pose estimation is crucial for diagnosing larynx diseases that can eventually lead to VF paralysis. The videoendoscopic examination used assess motility, usually estimating the change in anterior glottic angle (AGA). This a subjective and time-consuming procedure requiring extensive expertise. research proposes deep learning framework estimate from laryngoscopy frames acquired actual clinical practice. performs heatmap regression relying on three anatomically relevant keypoints as prior AGA computation, which estimated coordinates of predicted points. assessment proposed performed using newly collected dataset 471 124 patients, 28 whom with cancer. was tested various configurations compared other state-of-the-art approaches (direct glottal segmentation) both estimation, evaluation. obtained lowest root mean square error (RMSE) computed all (5.09, 6.56, 6.40 pixels, respectively) among models estimation. Also evaluation, reached average (MAE) ( $$5.87^{\circ }$$ 5 . 87 ∘ ). Results show allows perform small error, overcoming drawbacks algorithms, especially challenging images such pathologic subjects, presence noise, occlusion.

Language: Английский

Enhancing repeatability of follicle counting with deep learning reconstruction high-resolution MRI in PCOS patients DOI Creative Commons
Renjie Yang,

Yujie Zou,

Liang Li

et al.

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: Jan. 7, 2025

Abstract Follicle count, a pivotal metric in the adjunct diagnosis of polycystic ovary syndrome (PCOS), is often underestimated when assessed via transvaginal ultrasonography compared to MRI. Nevertheless, repeatability follicle counting using traditional MR images still compromised by motion artifacts or inadequate spatial resolution. In this prospective study involving 22 PCOS patients, we employed periodically rotated overlapping parallel lines with enhanced reconstruction (PROPELLER) and single-shot fast spin-echo (SSFSE) T2-weighted sequences suppress high-resolution ovarian Additionally, deep learning (DL) was utilized compensate noise SSFSE imaging. We performance DL (SSFSE-DL) conventional (SSFSE-C) PROPELLER detection, employing qualitative indices (blurring artifacts, subjective noise, conspicuity follicles) number per (FNPO) assessment. Despite similar between SSFSE-DL as one observer, outperformed SSFSE-C across all three indices, resulting FNPO These results highlighted potential imaging more dependable method for identifying ovary, thus facilitating accurate future clinical practices.

Language: Английский

Citations

0
Clinical feasibility of deep learning-driven magnetic resonance angiography collateral map in acute anterior circulation ischemic stroke DOI Creative Commons
YoungSook Jeon, Hong Gee Roh, Sumin Jung

et al.

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: Jan. 17, 2025

To validate the clinical feasibility of deep learning-driven magnetic resonance angiography (DL-driven MRA) collateral map in acute ischemic stroke. We employed a 3D multitask regression and ordinal neural network, called as 3D-MROD-Net, to generate DL-driven MRA maps. Two raters graded perfusion scores both conventional maps measured grading time. They also qualitatively assessed image quality Interrater inter-method agreements for between two were analyzed, along with comparison time quality. In analysis 296 stroke patients, agreement was almost perfect (κ = 0.96, 95% CI: 0.95-0.98). Compared maps, taken on shorter (P < 0.001 rater 1 P 0.003 2), superior 0.002 2). The demonstrates stroke, added benefits reduced generation interpretation time, improved map.

Language: Английский

Citations

0
Ultra-high-resolution brain MRI at 0.55T: bSTAR and its application to magnetization transfer ratio imaging DOI Creative Commons
Grzegorz Bauman, Roya Afshari, Oliver Bieri

et al.

Zeitschrift für Medizinische Physik, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 1, 2025

Language: Английский

Citations

0
Nonlinear Harmonics: A Gateway to Enhanced Image Contrast and Material Discrimination DOI Creative Commons
Pardis Biglarbeigi, Gourav Bhattacharya, Dewar Finlay

et al.

Advanced Science, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 28, 2025

Abstract Recent advancements in atomic force microscopy (AFM) have enabled detailed exploration of materials at the molecular and levels. These developments, however, pose a challenge: data generated by microscopic spectroscopic experiments are increasing rapidly both size complexity. Extracting meaningful physical insights from these datasets is challenging, particularly for multilayer heterogeneous nanoscale structures. In this paper, an unsupervised approach presented to enhance AFM image contrast analyzing nonlinear response cantilever interacting with material's surface using wavelet‐based AFM. This method simultaneously measures different frequencies harmonics single scan, without need additional hardware exciting multiple cantilevers' eigenmodes. developed enhancement (AFM‐ICE) employs learning, processing, fusion techniques. The applied interpret complex structures consist defects, deposited nanoparticles heterogeneities. Its substantial capability demonstrated improve differentiate between various components. methodology can pave way rapid precise determination material properties enhanced resolution.

Language: Английский

Citations

0
A Deep-Learning Approach for Vocal Fold Pose Estimation in Videoendoscopy DOI Creative Commons
Francesca Pia Villani, Maria Chiara Fiorentino, L. Fédérici

et al.

Deleted Journal, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 12, 2025

Abstract Accurate vocal fold (VF) pose estimation is crucial for diagnosing larynx diseases that can eventually lead to VF paralysis. The videoendoscopic examination used assess motility, usually estimating the change in anterior glottic angle (AGA). This a subjective and time-consuming procedure requiring extensive expertise. research proposes deep learning framework estimate from laryngoscopy frames acquired actual clinical practice. performs heatmap regression relying on three anatomically relevant keypoints as prior AGA computation, which estimated coordinates of predicted points. assessment proposed performed using newly collected dataset 471 124 patients, 28 whom with cancer. was tested various configurations compared other state-of-the-art approaches (direct glottal segmentation) both estimation, evaluation. obtained lowest root mean square error (RMSE) computed all (5.09, 6.56, 6.40 pixels, respectively) among models estimation. Also evaluation, reached average (MAE) ( $$5.87^{\circ }$$ 5 . 87 ∘ ). Results show allows perform small error, overcoming drawbacks algorithms, especially challenging images such pathologic subjects, presence noise, occlusion.

Language: Английский

Citations

0