Dataset Information

A hybrid mask RCNN-based tool to localize dental cavities from real-time mixed photographic images.

ABSTRACT: Nearly 3.5 billion humans have oral health issues, including dental caries, which requires dentist-patient exposure in oral examinations. The automated approaches identify and locate carious regions from dental images by localizing and processing either colored photographs or X-ray images taken via specialized dental photography cameras. The dentists' interpretation of carious regions is difficult since the detected regions are masked using solid coloring and limited to a particular dental image type. The software-based automated tools to localize caries from dental images taken via ordinary cameras requires further investigation. This research provided a mixed dataset of dental photographic (colored or X-ray) images, instantiated a deep learning approach to enhance the existing dental image carious regions' localization procedure, and implemented a full-fledged tool to present carious regions via simple dental images automatically. The instantiation mainly exploits the mixed dataset of dental images (colored photographs or X-rays) collected from multiple sources and pre-trained hybrid Mask RCNN to localize dental carious regions. The evaluations performed by the dentists showed that the correctness of annotated datasets is up to 96%, and the accuracy of the proposed system is between 78% and 92%. Moreover, the system achieved the overall satisfaction level of dentists above 80%.

SUBMITTER: Rashid U

PROVIDER: S-EPMC9044255 | biostudies-literature | 2022

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

A hybrid mask RCNN-based tool to localize dental cavities from real-time mixed photographic images.

Rashid Umer U Rashid Umer U Javid Aiman A Khan Abdur Rehman AR Liu Leo L Ahmed Adeel A Khalid Osman O Saleem Khalid K Meraj Shaista S Iqbal Uzair U Nawaz Raheel R

PeerJ. Computer science 20220218

Nearly 3.5 billion humans have oral health issues, including dental caries, which requires dentist-patient exposure in oral examinations. The automated approaches identify and locate carious regions from dental images by localizing and processing either colored photographs or X-ray images taken <i>via</i> specialized dental photography cameras. The dentists' interpretation of carious regions is difficult since the detected regions are masked using solid coloring and limited to a particular denta ...[more]

PMID: 35494840

Similar Datasets

Project description:BackgroundThe accurate detection of eyelid tumors is essential for effective treatment, but it can be challenging due to small and unevenly distributed lesions surrounded by irrelevant noise. Moreover, early symptoms of eyelid tumors are atypical, and some categories of eyelid tumors exhibit similar color and texture features, making it difficult to distinguish between benign and malignant eyelid tumors, particularly for ophthalmologists with limited clinical experience.MethodsWe propose a hybrid model, HM_ADET, for automatic detection of eyelid tumors, including YOLOv7_CNFG to locate eyelid tumors and vision transformer (ViT) to classify benign and malignant eyelid tumors. First, the ConvNeXt module with an inverted bottleneck layer in the backbone of YOLOv7_CNFG is employed to prevent information loss of small eyelid tumors. Then, the flexible rectified linear unit (FReLU) is applied to capture multi-scale features such as texture, edge, and shape, thereby improving the localization accuracy of eyelid tumors. In addition, considering the geometric center and area difference between the predicted box (PB) and the ground truth box (GT), the GIoU_loss was utilized to handle cases of eyelid tumors with varying shapes and irregular boundaries. Finally, the multi-head attention (MHA) module is applied in ViT to extract discriminative features of eyelid tumors for benign and malignant classification.ResultsExperimental results demonstrate that the HM_ADET model achieves excellent performance in the detection of eyelid tumors. In specific, YOLOv7_CNFG outperforms YOLOv7, with AP increasing from 0.763 to 0.893 on the internal test set and from 0.647 to 0.765 on the external test set. ViT achieves AUCs of 0.945 (95% CI 0.894-0.981) and 0.915 (95% CI 0.860-0.955) for the classification of benign and malignant tumors on the internal and external test sets, respectively.ConclusionsOur study provides a promising strategy for the automatic diagnosis of eyelid tumors, which could potentially improve patient outcomes and reduce healthcare costs.

Project description:In order to enhance the mitigation of crack occurrence and propagation within basement concrete structures, this research endeavors to propose an optimization methodology grounded in the Mask Region-based Convolutional Neural Network (Mask-RCNN) and an analysis of temperature effects. Initially, the Mask-RCNN algorithm is employed to perform image segmentation of the basement concrete structure, facilitating the precise identification of crack locations and shapes within the structure. Subsequently, the finite element analysis method is harnessed to simulate the structural stress and deformation in response to temperature variations. An optimization algorithm is introduced to adjust geometric parameters and material properties using insights from the temperature effect analysis. This algorithm aims to minimize stress concentration and deformation within the structure, thus diminishing the incidence and proliferation of cracks. In order to assess the efficacy of the optimization approach, an authentic basement concrete structure is selected for scrutiny, and the structure is monitored in real-time through the installation of strain gauges and monitoring equipment. These instruments track structural stress and deformation under diverse temperature conditions, and the evolution of cracks is meticulously documented. The outcomes demonstrate that by adjusting the structural geometric parameters and material properties, the crack density experiences a notable reduction of 60.22%. Moreover, the average crack length and width witness reductions of 40.24% and 35.43%, respectively, thereby corroborating the efficacy of the optimization method. Furthermore, an assessment of stress concentration and deformation within the structure is conducted. Through the optimization process, the maximum stress concentration in the structure diminishes by 25.22%, while the maximum deformation is curtailed by 30.32%. These results signify a substantial enhancement in structural stability. It is evident that the optimization algorithm exhibits robustness and stability in the context of crack control, consistently delivering favorable outcomes across diverse parameter configurations.

Project description:As with faces, participants are better at discriminating upright bodies than inverted bodies. This inversion effect is reliable for whole figures, namely, bodies with heads, but it is less reliable for headless bodies. This suggests that removal of the head disrupts typical processing of human figures, and raises questions about the role of faces in efficient body discrimination. In most studies, faces are occluded, but the aim here was to exclude faces in a more ecologically valid way by presenting photographic images of human figures from behind (about-facing), as well as measuring gaze to different parts of the figures. Participants determined whether pairs of sequentially presented body postures were the same or different for whole and headless figures. Presenting about-facing figures (heads seen from behind) and forward-facing figures with faces enabled a comparison of the effect of the presence or absence of faces. Replicating previous findings, there were inversion effects for forward-facing whole figures, but less reliable effects for headless images. There were also inversion effects for about-facing whole figures, but not about-facing headless figures. Accuracy was higher in the forward- compared to the about-facing conditions, but proportional dwell time was greater to bodies in about-facing images. Likewise, despite better discrimination of forward-facing upright compared to inverted whole figures, participants focused more on the heads and less on the bodies in upright compared to inverted images. However, there was no clear relationship between performance and dwell time proportions to heads. Body inversion effects (BIEs) were found with about-facing whole figures and headless forward-facing figures, despite the absence of faces. With inverted whole figures, there was a significant relationship between performance and greater looking at bodies, and less at heads suggesting that in more difficult conditions a focus on bodies is associated with better discrimination. Overall, the findings suggest that the visual system has greater sensitivity to bodies in their most experienced form, which is typically upright and with a head. Otherwise, the more a face is implied by the context, as in whole figures or forward- rather than about-facing headless bodies, the better the performance as holistic/configural processing is likely stronger.

Dataset Information

A hybrid mask RCNN-based tool to localize dental cavities from real-time mixed photographic images.

Publications

A hybrid mask RCNN-based tool to localize dental cavities from real-time mixed photographic images.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets