Search

Breast cancer is a common cancer that hits women causing thousands of casualties every year. A cancerous tumor causes an increase of temperature near the region of the tumor. The heat generated by the temperature transferred to the skin surface. The temperature in the tumor area is warmer than in the healthy area. Detecting breast cancer in early stages can save women's lives and lower the burden on the cost. Thermography is an imaging technique used for breast cancer detection. A dynamic thermography technique which is used to generate infrared images over a fixed time measured in minutes to detect the difference between the normal and cancerous areas in images. In this research, we propose a methodology to deal with the changes of temperature in patient's breasts by defining a set of efficient features resulted from extraction and reduction of coefficients obtained from breast thermogram images followed by classification. Texture feature methods (Histogram of Oriented Gradients (HOG) and Discrete Curvelet transform) are applied separately using the HH (high-high) and HL (high-low) sub band images of Discrete Wavelet transform (DWT). HOG-based features and Curvelet features are extracted by reducing coefficients' vectors returned by the two methods. Finally, Support Vector Machine (SVM) binary classifier is used to classify the images to either normal or abnormal. The proposed work has successfully achieved an Accuracy of 98.2%, Sensitivity of 97.7%, and Specificity of 98.2% through empirical studies using dynamic breast thermogram dataset. [ABSTRACT FROM AUTHOR]

Synchrotron radiation light source facilities are leading the way to ultrahigh resolution X-ray imaging. High resolution imaging is essential to understanding the fundamental structure and interaction of materials at the smallest length scale possible. Diffraction based methods achieve nanoscale imaging by replacing traditional objective lenses by pixelated area detectors and computational image reconstruction. Among these methods, ptychography is quickly becoming the standard for sub-30 nanometer imaging of extended samples, but at the expense of increasingly high data rates and volumes. This paper presents a new distributed algorithm for solving the ptychographic image reconstruction problem based on automatic differentiation. Input datasets are subdivided between multiple graphics processing units (GPUs); each subset of the problem is then solved either entirely independent of other subsets (asynchronously) or through sharing gradient information with other GPUs (synchronously). The algorithm was evaluated on simulated and real data acquired at the Advanced Photon Source, scaling up to 192 GPUs. The synchronous variant of our method outperformed an existing multi-GPU implementation in terms of accuracy while running at a comparable execution time. [ABSTRACT FROM AUTHOR]

In railway operations, delays are used as one of most important factors to quantify and evaluate the quality of the railway services. However, data about stochastic disturbances and the causes of the delays are hard to be collected and measured. The efforts to manually estimate these disturbances are also considerably high. In this paper, a method for the automatic calibration of disturbance parameters, which are used to generate stochastic disturbances in simulation tools, is developed with the support of the reinforcement learning technique. Simulation and application results show that the efforts for calibrating parameters can be significantly reduced with ensured consistency between simulation models and the actual railway operations. [ABSTRACT FROM AUTHOR]

A problem of specification of motion parameters of a spacecraft (SC) for remote Earth probing by results of measurements of navigation parameters is considered. We analyze specific features of ballistic-navigation support for a single-point control technology. It is demonstrated, that this technology does not satisfy requirements for accuracy and efficiency of obtaining necessary ballistic-navigation information. We suggest enhancing accuracy and efficiency of calculation of ballistic information by minimization of the goal function of specification of motion parameters of an SC by search methods. Results of analysis of the developed algorithm of specification of motion parameters of an SC on the basis of the simplex optimization method are presented. [ABSTRACT FROM AUTHOR]

Stochastic gradient descent (SGD) algorithm and its variations have been effectively used to optimize neural network models. However, with the rapid growth of big data and deep learning, SGD is no longer the most suitable choice due to its natural behavior of sequential optimization of the error function. This has led to the development of parallel SGD algorithms, such as asynchronous SGD (ASGD) and synchronous SGD (SSGD) to train deep neural networks. However, it introduces a high variance due to the delay in parameter (weight) update. We address this delay in our proposed algorithm and try to minimize its impact. We employed guided SGD (gSGD) that encourages consistent examples to steer the convergence by compensating the unpredictable deviation caused by the delay. Its convergence rate is also similar to A/SSGD, however, some additional (parallel) processing is required to compensate for the delay. The experimental results demonstrate that our proposed approach has been able to mitigate the impact of delay for the quality of classification accuracy. The guided approach with SSGD clearly outperforms sequential SGD and even achieves an accuracy close to sequential SGD for some benchmark datasets. • Its convergence rate of O 1 ρ T + σ 2 shows its applicability for the real-time systems. • The proposed method outperforms synchronous/asynchronous SGD. • The proposed method is compatible with other variations of SGD such as RMSprop. • The delay in parameter updates happens due to several gradients computation in parallel. [ABSTRACT FROM AUTHOR]