Abstract
Medical image classification plays an important role in disease diagnosis since it can provide important reference information for doctors. The supervised convolutional neural networks (CNNs) such as DenseNet provide the versatile and effective method for medical image classification tasks, but they require large amounts of data with labels and involve complex and time-consuming training process. The unsupervised CNNs such as principal component analysis network (PCANet) need no labels for training but cannot provide desirable classification accuracy. To realize the accurate medical image classification in the case of a small training dataset, we have proposed a light-weighted hybrid neural network which consists of a modified PCANet cascaded with a simplified DenseNet. The modified PCANet has two stages, in which the network produces the effective feature maps at each stage by convoluting inputs with various learned kernels. The following simplified DenseNet with a small number of weights will take all feature maps produced by the PCANet as inputs and employ the dense shortcut connections to realize accurate medical image classification. To appreciate the performance of the proposed method, some experiments have been done on mammography and osteosarcoma histology images. Experimental results show that the proposed hybrid neural network is easy to train and it outperforms such popular CNN models as PCANet, ResNet and DenseNet in terms of classification accuracy, sensitivity and specificity.
Highlights
During the process of disease diagnosis, doctors need to exam such numerous medical images as X-ray images, magnetic resonance (MR) images and ultrasound images
The modified principal component analysis network (PCANet) is utilized because its good feature extraction ability paves the way for the following network, and its unsupervised and interpretable learning strategy can alleviate the requirement for labeled data and render the hybrid neural network more reliable than the regular convolutional neural networks (CNNs) in medical image classification in the case of a small number of training samples
To demonstrate the performance of our proposed network, comparisons are made among the popular convolution neural networks such as VGG, ResNet-50, DenseNet-121, the original PCANet, our simplified DenseNet without the modified PCANet as the input and the proposed hybrid neural network (HybridNet) operating on the Digital Database for Screening Mammography (DDSM) dataset, the osteosarcoma histology images [46], [47] and the mammographic image analysis society (MIAS) dataset [48]
Summary
During the process of disease diagnosis, doctors need to exam such numerous medical images as X-ray images, magnetic resonance (MR) images and ultrasound images. We have designed a light-weighted hybrid neural network for medical image classification by combining a modified PCANet and a simplified DenseNet model. The hybrid neural network utilizes a modified two-stage PCANet to extract the low-level features and applies a simplified DenseNet model to extracting the high-level features for accurate medical image classification. The modified PCANet is utilized because its good feature extraction ability paves the way for the following network, and its unsupervised and interpretable learning strategy can alleviate the requirement for labeled data and render the hybrid neural network more reliable than the regular CNNs in medical image classification in the case of a small number of training samples. In the hybrid neural network, the PCANet and the DenseNet will be trained in order using two different strategies
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
