Utilizing Mask R-CNN for Detection and Segmentation of Oral Diseases

Anantharaman, R; Velazquez, Matthew; Lee, Yugyung

doi:10.1109/bibm.2018.8621112

Cited by 98 publications

(49 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Methods related to the automated diagnosis of oral cancer, OPMDs and benign lesions are largely based on microscopic images [9]- [12]. Other literature covers the use of multidimensional hyperspectral images of the mouth [13], the use of CT (computed tomography) images [14], the use of autofluorescence [15], [16] and fluorescence imaging [17] which focused on relative close-ups of the oral lesions and, finally, standard white light images which captured oral cavity structures [18]- [20].…”

Section: Introductionmentioning

confidence: 99%

“…Early publications in the field focused on texture based features, Thomas [18] used the grey level co-occurrence matrix and grey level run-length, whilst Krishnan [9] made use of higher order spectra, local binary pattern and laws texture energy. The more recent papers [10]- [17], [19], [20] have made the shift towards employing deep learning, which are artificial neural networks that consist of many layers of neurons and rely on large datasets and fast computing power to enable them to learn complex patterns. More specifically these publications made use of the deep convolutional neural network (CNN) whose architectures made the explicit assumption that the inputs were in the form of images.…”

Section: Introductionmentioning

confidence: 99%

“…Object detection frameworks have been explored in the medical imaging domain, with the Faster R-CNN being applied to colon polyp detection [31] and detection and classification of lesions on mammograms [32]. Anantharaman [20] applied the Mask R-CNN to oral images using a dataset of 40 images, to detect the benign oral lesions of cold sores (herpes labialis) and canker sores (aphthous ulcers). Their evaluation was based on instance segmentation as opposed to bounding box detection.…”

Section: Introductionmentioning

confidence: 99%

“…The former made use of ResNet-101 [33] which was a powerful CNN. The latter built on other work [20], but used the Faster R-CNN rather than the Mask R-CNN in order to focus on bounding box detection performance and, more importantly, this framework was now applied to oral cancer and OPMDs. Fig.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Automated Detection and Classification of Oral Lesions Using Deep Learning for Early Detection of Oral Cancer

et al. 2020

View full text Add to dashboard Cite

Oral cancer is a major global health issue accounting for 177,384 deaths in 2018 and it is most prevalent in low-and middle-income countries. Enabling automation in the identification of potentially malignant and malignant lesions in the oral cavity would potentially lead to low-cost and early diagnosis of the disease. Building a large library of well-annotated oral lesions is key. As part of the MeMoSA ® (Mobile Mouth Screening Anywhere) project, images are currently in the process of being gathered from clinical experts from across the world, who have been provided with an annotation tool to produce rich labels. A novel strategy to combine bounding box annotations from multiple clinicians is provided in this paper. Further to this, deep neural networks were used to build automated systems, in which complex patterns were derived for tackling this difficult task. Using the initial data gathered in this study, two deep learning based computer vision approaches were assessed for the automated detection and classification of oral lesions for the early detection of oral cancer, these were image classification with ResNet-101 and object detection with the Faster R-CNN. Image classification achieved an F1 score of 87.07% for identification of images that contained lesions and 78.30% for the identification of images that required referral. Object detection achieved an F1 score of 41.18% for the detection of lesions that required referral. Further performances are reported with respect to classifying according to the type of referral decision. Our initial results demonstrate deep learning has the potential to tackle this challenging task. INDEX TERMS Composite annotation, deep learning, image classification, object detection, oral cancer, oral potentially malignant disorders.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Automated Detection and Classification of Oral Lesions Using Deep Learning for Early Detection of Oral Cancer

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Esta implementación parte del framework realizado por Matterport Inc. [23] en un trabajo anterior, el cual se proporciona en código abierto bajo una licencia MIT. La arquitectura y el funcionamiento de esta red se encuentra detallado en [9] [24]. El primer bloque de la arquitectura está formado por tres sub-bloques: Backbone que extrae mapas de características de las imágenes con capas convolucionales (normalmente ResNet50 o ResNet101), Feature Pyramid Network (FPN) [25] que proporciona que las características de alto nivel estén a diferentes escalas y con información de diferentes niveles, y Region Proposal Network (RPN) introducido en [10].…”

Section: Bases Del Modelo Neuronalunclassified

Generación de datos sintéticos con objetos de cocina para entrenar redes neuronales de convolución

Benages-Pardo¹,

Sagüés-Tanco²,

López‐Nicolás³

et al. 2020

XL Jornadas De Automática: Libro De Actas (Ferrol, 4-6 De Septiembre De 2019)

View full text Add to dashboard Cite

El reconocimiento de objetos ha sido un problema ampliamente analizado en el campo de la visión por computador desde hace muchos años. En este artículo utilizamos el modelo de red neuronal Mask R-CNN para detectar y segmentar algunos objetos de cocinas usando imágenes RGB. Existen algunas bases de datos públicas con los objetos segmentados y etiquetados por clases, para entrenar este tipo de redes pero normalmente no incluyen clases para aplicaciones concretas. Crear un conjunto de datos etiquetado a mano completamente es una tarea muy tediosa y larga. Proponemos un método eficiente para crear imágenes etiquetadas con poco esfuerzo combinando imágenes de fondos con objetos segmentados. Las imágenes se pueden crear manteniendo el realismo en cuanto a la posición y la escala de los objetos, o automáticamente mediante un posicionamiento aleatorio. Se presenta finalmente una comparación de los resultados obtenidos al entrenar la red con ambos conjuntos de datos creados de forma sintética.

show abstract

Improving lung cancer detection using faster region‐based convolutional neural network aided with fuzzy butterfly optimization algorithm

Sinthia

Malathi²,

Anitha

et al. 2022

Concurrency and Computation

View full text Add to dashboard Cite

Lung cancer is the most deadly type of cancer and it is caused by genetic variations in lung tissues. Other causes of lung cancer are alcohol, smoking, and threatening gas exposures. The diagnosis of lung cancer is an intricate task and early detection of lung cancer can help to get the exact treatment in advance. The application of a computer-aided diagnosis process helps to predict lung cancer earlier, nonetheless, it does not provide better accuracy. The overfitting nature of features and dimensionality of lung cancer can prevent it from obtaining maximum accuracy. Hence, we proposed a novel faster region convolutional neural network (RCNN) based fuzzy butterfly optimization algorithm (FBOA) to achieve better prediction accuracy and effectiveness. The proposed Faster RCNN can provide better positioning of lung cancer swiftly and effectively and the FBOA approach can be used to perform two-stage classification. The fuzzy rules used in the FBOA can be utilized to find the severity of the lung cancer and differentiate the Benign stage and Malignant stage effectively.The experimental analyses are performed in MATLAB simulation. The preprocessing of images is performed by different tools from MATLAB and is to format the images as required. The cancer imaging archive (TCIA) dataset is utilized to analyze the performance of the proposed method and compared with various state-of-art works. The performance of the proposed method is evaluated by using different evaluation metrics such as precision, recall, F-measure, and accuracy and attained 99, 98, 99, and 97% respectively. Thus, our proposed method outperforms all the other approaches.

show abstract

Utilizing Mask R-CNN for Detection and Segmentation of Oral Diseases

Cited by 98 publications

References 19 publications

Automated Detection and Classification of Oral Lesions Using Deep Learning for Early Detection of Oral Cancer

Automated Detection and Classification of Oral Lesions Using Deep Learning for Early Detection of Oral Cancer

Generación de datos sintéticos con objetos de cocina para entrenar redes neuronales de convolución

Improving lung cancer detection using faster region‐based convolutional neural network aided with fuzzy butterfly optimization algorithm

Contact Info

Product

Resources

About