Convolutional Neural Network Models for Content Based X-Ray Image Classification

Arti, P; Agrawal, Abhishek; Adishesh, A; Lahari, V M; Niranjana, Krupa B

doi:10.1109/wiecon-ece48653.2019.9019943

Cited by 4 publications

(2 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…CosFace [14] used the estimated clustering uncertainty of unlabelled samples to adjust the loss function weight to reduce the overlapping-identity label noise; however, it requires balanced labelled and unlabelled samples to estimate clustering uncertainty accurately, which is a major limitation. There are a few methods based on Sensors 2022, 22, 9967 2 of 12 transfer learning [15,16] and meta-learning [17,18]. Ahn et al [15] proposed a hierarchical unsupervised feature extractor, which has a convolutional autoencoder on top of a pretrained convolutional neural network (CNN).…”

Section: Introductionmentioning

confidence: 99%

“…Ahn et al [15] proposed a hierarchical unsupervised feature extractor, which has a convolutional autoencoder on top of a pretrained convolutional neural network (CNN). Arti Pet al [16] fine-tuned the pre-trained AlexNet [2] and GoogleNet [19] for X-ray image classification. Maicas et al [18] designed an unsupervised pretext task for meta-learning and then trained the model for medical image classification.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Medical Image Classification Based on Semi-Supervised Generative Adversarial Network and Pseudo-Labelling

Liu

Ning

Liu

2022

Sensors

View full text Add to dashboard Cite

Deep learning has substantially improved the state-of-the-art in object detection and image classification. Deep learning usually requires large-scale labelled datasets to train the models; however, due to the restrictions in medical data sharing and accessibility and the expensive labelling cost, the application of deep learning in medical image classification has been dramatically hindered. In this study, we propose a novel method that leverages semi-supervised adversarial learning and pseudo-labelling to incorporate the unlabelled images in model learning. We validate the proposed method on two public databases, including ChestX-ray14 for lung disease classification and BreakHis for breast cancer histopathological image diagnosis. The results show that our method achieved highly effective performance with an accuracy of 93.15% while using only 30% of the labelled samples, which is comparable to the state-of-the-art accuracy for chest X-ray classification; it also outperformed the current methods in multi-class breast cancer histopathological image classification with a high accuracy of 96.87%.

show abstract