Multistage transfer learning technique for classifying rare medical datasets

Aswiga, R. V.; Aishwarya, R.; Shanthi, A. P.

doi:10.1007/s12652-021-02989-1

Cited by 7 publications

(3 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hence, this becomes the tedious work for extracting and describing features automatically during the training phase, due to which the classi cation error is minimized when the model is trained. As mentioned in the work proposed by [18]- [20], the concept of advanced deep learning techniques have been employed for designing CNN framework for medical domain. The existing CNN model for the medical application consists of four types of layers namely, the convolutional layer, which is the rst layer; its purpose is to identify the essential features received as input.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

An automatic detection and classification of diabetes mellitus using CNN

Aswiga

Karpagam

et al. 2023

Soft Comput

Self Cite

View full text Add to dashboard Cite

Machine learning algorithms have recently attracted more attention in the medical eld and have been successfully used in a variety of medical applications. Machine learning is used to assist medical practitioners to solve complicated problems. Thus to function autonomously without seeking any assistance or help from a person, a well-known branch of machine learning called deep learning completes several related tasks parallely. It is successfully used in a variety of applications, including disease prediction and disease progression. Since tasks are assumed to be related to one another, existing learning methods adapts the Convolution neural network (CNN) to classify the diabetes mellitus disease effectively. To detect the abnormality, the proposed work uses a deep learning network that includes a novel CNN model that splits the data into separate training and testing sets before doing classi cation. Using CNN, the proposed work achieves better accuracy and outperforms well to the best of our knowledge.

show abstract

Section: Methodsmentioning

confidence: 99%

“…The functionalities of each and every layer are described below. Inspired by the existing work [18]- [20], the proposed work also focuses on developing a CNN model for processing diabetes dataset. The architecture diagram of proposed framework is presented below.…”

Section: Methodsmentioning

confidence: 99%

An automatic detection and classification of diabetes mellitus using CNN

Aswiga

Karpagam

et al. 2023

Soft Comput

Self Cite

View full text Add to dashboard Cite

show abstract

“…Their dataset originally comprised 1089 images and 12 classes, and after augmentation and transfer learning, the model showed a top accuracy of 97.66% 16 . Furthermore, AI has been applied to small datasets to detect abnormalities, such as tumors 17 , 18 , has demonstrated efficiency in supporting diagnosis from radiographs. The ensemble method was also applied to improve the accuracy 19 , which has better performance than a single classifier 20 .…”

Section: Introductionmentioning

confidence: 99%

Detection of incomplete atypical femoral fracture on anteroposterior radiographs via explainable artificial intelligence

Kim

Moon

Goh

et al. 2023

Sci Rep

View full text Add to dashboard Cite

One of the key aspects of the diagnosis and treatment of atypical femoral fractures is the early detection of incomplete fractures and the prevention of their progression to complete fractures. However, an incomplete atypical femoral fracture can be misdiagnosed as a normal lesion by both primary care physicians and orthopedic surgeons; expert consultation is needed for accurate diagnosis. To overcome this limitation, we developed a transfer learning-based ensemble model to detect and localize fractures. A total of 1050 radiographs, including 100 incomplete fractures, were preprocessed by applying a Sobel filter. Six models (EfficientNet B5, B6, B7, DenseNet 121, MobileNet V1, and V2) were selected for transfer learning. We then composed two ensemble models; the first was based on the three models having the highest accuracy, and the second was based on the five models having the highest accuracy. The area under the curve (AUC) of the case that used the three most accurate models was the highest at 0.998. This study demonstrates that an ensemble of transfer-learning-based models can accurately classify and detect fractures, even in an imbalanced dataset. This artificial intelligence (AI)-assisted diagnostic application could support decision-making and reduce the workload of clinicians with its high speed and accuracy.

show abstract

A Multilevel Transfer Learning Technique and LSTM Framework for Generating Medical Captions for Limited CT and DBT Images

Aswiga

Shanthi

2022

J Digit Imaging

View full text Add to dashboard Cite

Multistage transfer learning technique for classifying rare medical datasets

Cited by 7 publications

References 33 publications

An automatic detection and classification of diabetes mellitus using CNN

An automatic detection and classification of diabetes mellitus using CNN

Detection of incomplete atypical femoral fracture on anteroposterior radiographs via explainable artificial intelligence

A Multilevel Transfer Learning Technique and LSTM Framework for Generating Medical Captions for Limited CT and DBT Images

Contact Info

Product

Resources

About