A dataset of laryngeal endoscopic images with comparative study on convolution neural network-based semantic segmentation

Abstract

Purpose Automated segmentation of anatomical structures in medical image analysis is a prerequisite for autonomous diagnosis as well as various computer and robot aided interventions. Recent methods based on deep convolutional neural networks (CNN) have outperformed former heuristic methods. However, those methods were primarily evaluated on rigid, real-world environments. In this study, existing segmentation methods were evaluated for their use on a new dataset of transoral endoscopic exploration. Methods Four machine learning based methods SegNet, UNet, ENet and ErfNet were trained with supervision on a novel 7-class dataset of the human larynx. The dataset contains 536 manually segmented images from two patients during laser incisions. The Intersection-over-Union (IoU) evaluation metric was used to measure the accuracy of each method. Data augmentation and network ensembling were employed to increase segmentation accuracy. Stochastic inference was used to show uncertainties of the individual models. Patient-to-patient transfer was investigated using patient-specific fine-tuning. Results In this study, a weighted average ensemble network of UNet and ErfNet was best suited for the segmentation of laryngeal soft tissue with a mean IoU of 84.7 %. The highest efficiency was achieved by ENet with a mean inference time of 9.22 ms per image. It is shown that 10 additional images from a new patient are sufficient for patient-specific fine-tuning. Conclusion CNN-based methods for semantic segmentation are applicable to endoscopic images of laryngeal soft tissue. The segmentation can be used for active constraints or to monitor morphological changes and autonomously detect pathologies. Further improvements could be achieved by using a larger dataset or training the models in a self-supervised manner on additional unlabeled data.

Efficient Biomedical Image Segmentation on EdgeTPUs at Point of Care

“…segmenter variability common to glottis segmentation [57], and thus, are performing as good as experts in the field. Our method is also superior to other (fully automatic) glottis segmentation algorithms available [8], [9], [12], as it is validated on a diverse dataset, computationally highly efficient and also highly portable. However, by constraining the network capacity, the network may generalize less and thus, produce potentially erroneous segmentations depending on the input data.…”

BAGLS, a multihospital Benchmark for Automatic Glottis Segmentation

“…Trained experts anecdotally require about 15 minutes to segment a 1,000 frames long HSV recording using specifically developed software. Therefore, several previous works have explored the possibility of performing an automated segmentation of the glottal area 16,17,[43][44][45][46] .…”

Fully automatic segmentation of glottis and vocal folds in endoscopic laryngeal high-speed videos using a deep Convolutional LSTM Network

“…Although the application of deep CNNs has achieved highly reliable results for various semantic segmentation tasks on medical images, CNNs have hardly been investigated on their suitability for the analysis of laryngeal high-speed videos. Up to the present only very few publications dealing with the segmentation of laryngeal structures using artificial Neural Networks can be found in literature [72][73][74]. The few presented approaches have in common that the images of video sequences are processed independently from each other.…”