“…If an image has one or more contours associated with it, the same transformation is applied to the contours. Geometric transformations are so common that they were utilised by 92 of the 93 basic augmentation studies 15–106 …”
Summary
Research in artificial intelligence for radiology and radiotherapy has recently become increasingly reliant on the use of deep learning‐based algorithms. While the performance of the models which these algorithms produce can significantly outperform more traditional machine learning methods, they do rely on larger datasets being available for training. To address this issue, data augmentation has become a popular method for increasing the size of a training dataset, particularly in fields where large datasets aren’t typically available, which is often the case when working with medical images. Data augmentation aims to generate additional data which is used to train the model and has been shown to improve performance when validated on a separate unseen dataset. This approach has become commonplace so to help understand the types of data augmentation techniques used in state‐of‐the‐art deep learning models, we conducted a systematic review of the literature where data augmentation was utilised on medical images (limited to CT and MRI) to train a deep learning model. Articles were categorised into basic, deformable, deep learning or other data augmentation techniques. As artificial intelligence models trained using augmented data make their way into the clinic, this review aims to give an insight to these techniques and confidence in the validity of the models produced.
“…If an image has one or more contours associated with it, the same transformation is applied to the contours. Geometric transformations are so common that they were utilised by 92 of the 93 basic augmentation studies 15–106 …”
Summary
Research in artificial intelligence for radiology and radiotherapy has recently become increasingly reliant on the use of deep learning‐based algorithms. While the performance of the models which these algorithms produce can significantly outperform more traditional machine learning methods, they do rely on larger datasets being available for training. To address this issue, data augmentation has become a popular method for increasing the size of a training dataset, particularly in fields where large datasets aren’t typically available, which is often the case when working with medical images. Data augmentation aims to generate additional data which is used to train the model and has been shown to improve performance when validated on a separate unseen dataset. This approach has become commonplace so to help understand the types of data augmentation techniques used in state‐of‐the‐art deep learning models, we conducted a systematic review of the literature where data augmentation was utilised on medical images (limited to CT and MRI) to train a deep learning model. Articles were categorised into basic, deformable, deep learning or other data augmentation techniques. As artificial intelligence models trained using augmented data make their way into the clinic, this review aims to give an insight to these techniques and confidence in the validity of the models produced.
“…For feature extraction, a pre-trained ResNet 50 model is used. This model can handle the gradient disappearance and degradation problem of general CNN by using the residual blocks ( Lu et al, 2020 ). The performance of CNN has improved by the depth of proposed residual blocks.…”
“…Basically, convolutional layer is composed of filters which have to be convolved across the width as well as the height of input data. Besides, the final outcome of a convolutional layer is obtained with the help of dot product over a filter weight content and exclusive image location 17 . It is considered a 2D activation map which provides immediate filter responses in a spatial location.…”
Section: Scene Classification Methodologymentioning
Summary
In present days, unmanned aerial vehicles (UAVs) have gained significant interest among researchers and academicians. The UAVs were found useful in diverse application areas, namely, intelligent transportation system, disaster management, surveillance, and wildlife monitoring. Clustering is a well‐known energy‐efficient technique, which elects a cluster head (CH) among other nodes. At the same time, scene classification from the high‐resolution remote sensing images captured by UAV is also a major issue in the UAV networks. In order to resolve these problems, this paper projects novel energy‐efficient cluster‐based UAV networks with deep learning (DL)‐based scene classification method. The proposed model involves a clustering with parameter tuned residual network (C‐PTRN) model, which operates on two major phases such as cluster construction and scene classification. Initially, the UAVs are clustered using the type II fuzzy logic (T2FL) technique on the basis of residual energy, distance to nearby UAVs, and UAV degree. Next, the chosen CHs transmit the captured images to the base station (BS). At the second level, a DL‐based ResNet50 technique is employed for scene classification. To tune the hyperparameters of the ResNet50 model, water wave optimization (WWO) algorithm is used. At last, kernel extreme learning machine (KELM) model is used to perform the scene classification process. In order to ensure the performance of the proposed method, a detailed set of simulations takes place under different dimensions. The obtained results ensured that the C‐PTRN model has showcased supreme outcome with the maximum precision of 95.89%, recall of 98.91%, and F score of 96.54%.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.