Fetal neurosonography is potentially used to examine the fetal brain by scanning the trans-thalamic (TT), trans-cerebellum (TC), and trans-ventricular (TV) planes. Cross-sectional analysis of these planes is useful to assess the brain anatomy, development, and abnormality for intervention and treatment plans even at the postnatal stage. To minimize the errors and processing time involved in the traditional manual subjective approach, the automatic classification of fetal brain planes is crucial. In this study, a deep learning-based method for automatically categorizing fetal brain planes from ultrasound images is proposed and evaluated. Firstly, the brain region has been segmented from the fetal brain ultrasound images using U-Net to prepare an efficient data set for the classifier model. Then, an ensemble convolutional neural network (CNN) model including well-known Inception V3, ResNet50-V2, and DenseNet-201 models with max voting is designed to classify the segmented brain planes. 2019 fetal brain ultrasound images from a widely used publicly accessible experts-annotated dataset are used to evaluate the performance of the proposed framework. The obtained results analysis shows that using the segmented images as input improves the performance of the classifier from its raw form. The gradient class activation mapping (Grad-CAM) based inspection shows noteworthy localization capability of the last convolution layer. The ensemble model has also outperformed its individual model’s performance. The suggested categorization framework is satisfactory compared to related recent works, with a testing accuracy of 97.68%. The proposed framework for fetal brain plane classification is expected to be useful for clinical applications.

Ultrasound based breast screening is gaining attention recently especially for dense breast. The technological advancement, cancer awareness, and cost-safety-availability benefits lead rapid rise of breast ultrasound market. The irregular shape, intensity variation, and additional blood vessels of malignant cancer are distinguishable in ultrasound images from the benign phase. However, classification of breast cancer using ultrasound images is a difficult process owing to speckle noise and complex textures of breast. In this paper, a breast cancer classification method is presented using VGG16 model based transfer learning approach. We have used median filter to despeckle the images. The layers for convolution process of the pretrained VGG16 model along with the maxpooling layers have been used as feature extractor and a proposed fully connected two layers deep neural network has been designed as classifier. Adam optimizer is used with learning rate of 0.001 and binary cross-entropy is chosen as the loss function for model optimization. Dropout of hidden layers is used to avoid overfitting. Breast Ultrasound images from two databases (total 897 images) have been combined to train, validate and test the performance and generalization strength of the classifier. Experimental results showed the training accuracy as 98.2% and testing accuracy as 91% for blind testing data with a reduced of computational complexity. Gradient class activation mapping (Grad-CAM) technique has been used to visualize and check the targeted regions localization effort at the final convolutional layer and found as noteworthy. The outcomes of this work might be useful for the clinical applications of breast cancer diagnosis.

Lung tumor is the result of abnormal and uncontrolled cell division and growth in lung region. Earlier detection and staging of lung tumor is of great importance to increase the survival rate of the suffered patients. In this paper, a fast and robust Fuzzy c-means clustering method is used for segmenting the tumor region from lung CT images. Morphological reconstruction process is performed prior to Fuzzy c-means clustering to achieve robustness against noises. The computational efficiency is improved through median filtering of membership partition. Tumor masks are then reconstructed using surface based and shape based filtering. Different features are extracted from the segmented tumor region including maximum diameter and the tumor stage is determined according to the tumor staging system of American Joint Commission on Cancer. 3D shape of the segmented tumor is reconstructed from series of 2D CT slices for volume measurement. The accuracy of the proposed system is found as 92.72% for 55 randomly selected images from the RIDER Lung CT dataset of Cancer imaging archive. Lower complexity in terms of iterations and connected components as well as better noise robustness are found in comparison with conventional Fuzzy c-means and k-means clustering techniques.