Search

Automated visual identification of Benign and Malignant breast tumors even with ultrasound (US) B-Mode image is still an open area of research. This paper presents parametric image-based approach to classify and detect benign and malignant breast tumors from ultrasound images using a custom-made convolutional neural network architecture. The Rician Inverse Gaussian (RiIG) distribution is presented here as a suitable model for describing the statistics of the ultrasound images in the Contourlet Transform domain. Locally computed values of the dispersion parameters of the RiIG distribution in various contourlet sub-bands yield parametric images those are classified using the proposed convolutional neural network. Experiments are conducted on a publicly available dataset of 250 images, of 100 belong to the benign Fibroadenoma and 150 in the malignant category. The proposed method provides accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive values (NPV) of 96%, 97.3%, 94.12%, 96% and 96%, respectively. It is also shown that the accuracy obtained by the Proposed Method is higher than several recently reported results.

The movement of a projectile, which can be a mass or a charge, through a static field, can be important in various fields of physics ranging from particle physics to rocket science. In many cases, the movement of the projectile can be solely dependent on the initial velocity provided to the projectile. The determination of this initial velocity can be very complicated if the existing static field is too complex. In such cases, neural networks can be the key to modelling the environment and accurately estimate the initial velocity. This AI-driven velocity estimation can be extremely difficult using supervised learning due to the lack of seen examples of the state (initial position) to action (velocity) relation, which is also known as the policy. In this work, this problem is treated as a continuous-action-space contextual bandit problem and a neural network-based actor critic method is used to model the environment and determine the action policy. This method is tested on two different static field configurations using a Python-based GUI named Pygame. After training, the success rates of the projectile hitting the targeted destinations are 94.8% and 77.9% respectively, which implies that the proposed method can successfully model the action policy with high accuracy.