Measurement of laryngeal elevation by automated segmentation using Mask R-CNN

Abstract

The methods of measuring laryngeal elevation during swallowing are time-consuming. We aimed to propose a quick-to-use neural network (NN) model for measuring laryngeal elevation quantitatively using anatomical structures auto-segmented by Mask region-based convolutional NN (R-CNN) in videofluoroscopic swallowing study. Twelve videofluoroscopic swallowing study video clips were collected. One researcher drew the anatomical structure, including the thyroid cartilage and vocal fold complex (TVC) on respective video frames. The dataset was split into 11 videos (4686 frames) for model development and one video (532 frames) for derived model testing. The validity of the trained model was evaluated using the intersection over the union. The mean intersections over union of the C1 spinous process and TVC were 0.73 ± 0.07 [0–0.88] and 0.43 ± 0.19 [0–0.79], respectively. The recall rates for the auto-segmentation of the TVC and C1 spinous process by the Mask R-CNN were 86.8% and 99.8%, respectively. Actual displacement of the larynx was calculated using the midpoint of the auto-segmented TVC and C1 spinous process and diagonal lengths of the C3 and C4 vertebral bodies on magnetic resonance imaging, which measured 35.1 mm. Mask R-CNN segmented the TVC with high accuracy. The proposed method measures laryngeal elevation using the midpoint of the TVC and C1 spinous process, auto-segmented by Mask R-CNN. Mask R-CNN auto-segmented the TVC with considerably high accuracy. Therefore, we can expect that the proposed method will quantitatively and quickly determine laryngeal elevation in clinical settings.

A revolutionary acute subdural hematoma detection based on two-tiered artificial intelligence model

“…It is known that the MASK R-CNN method was used in segmentation problems using medical images and gave successful results. [24][25][26][27][28] The elimination of areas labeled as hemorrhagic due to artifacts, as a result of the segmentation process performed at the first level, was achieved by the SVM algorithm used at the second level. If the average of the HU value was used alone, the model was more affected by the errors that may occur during the hemorrhagic area marking (marking the areas of the bone, cavity, artifact, etc.).…”

“…When the studies on the data compiled specifically for the problem are examined, it has been determined that the success achieved with these data is relatively low. [24][25][26][27][28] In the study of Arbabshirani et al, in which they classified the data containing different hemorrhage types with 3D images as having or absent hemorrhage, the success was evaluated in the area under the curve value, and the accuracy was obtained as 84.6%. [46] The data set used in the study consists of CT images that have not been used in the previous studies and are not open access.…”

Temporal Sequence of Laryngeal Vestibule Closure and Reopening is Associated With Airway Protection

“…Recent technological advancements enable noninvasive detection of swallowing kinematics solely based on swallowing sounds and vibrations. [41][42][43][44][45][46] Other computer vision and artificial intelligence techniques were applied for automated frame-by-frame hyoid 47 and laryngeal 48 analyses on VFSS. These extracted measurements can be incorporated to our temporal understanding of laryngeal kinematics to estimate the risk of penetration or aspiration.…”