Narrow Results

Although mammography is still the benchmark technique for breast cancer detection, many advantages of thermography make it a suitable adjunct tool for early detection. This paper describes the development of a computer-aided system for use together with thermography to assist in the detection and visualization/analysis of breast tumors. The system consists of a detection module for predicting the presence of tumors from thermograms, and a visualization module for generating the 3-D volumetric geometry of the suspected tumor inside the breast based on the 2-D thermogram. Detection is achieved through an artificial neural network taking the thermogram image as input, while the visualization is obtained by generating the 3-D model of the breast that produces a matching thermal image as the thermogram under a 3-D finite element analysis. A study with 200 subjects indicate that the detection sensitivity was good but the specificity was poor, but the reverse performance result was true for another back-propagation neural network which used physiological data instead of thermograms as input. This suggests that overall prediction capability can be improved by appropriate combination of the two results.

The treatment of early development of breast tumor has a higher success rate. This paper presents a framework for the early discovery of breast cancer. The objective is to assist the general practitioners and specialists in the detection of breast tumor. The proposed detection process consists of a preliminary screening process and a prediction process. The preliminary screening process using thermography aims to complement the detailed screening operation using mammography. The prediction process using artificial intelligence techniques aims to use past records of other similar cases to enhance the forecast of breast cancer development. The paper discusses the issues and techniques for the implementation of the proposed framework. These include the preliminary screening process, the retrieval of the relevant cases, and the prediction of the risk of developing breast cancer based on the thermographs, environmental/social data, physiological information, genetic factors, and medical records. This work constitutes initial effort to lessen the burden of medical professionals and increase the chances of successful treatment for patients in the fight against breast cancer.

Magnetic resonance imaging (MRI) plays an integral role among the advanced techniques for detecting a brain tumor. The early detection of brain tumor with proper automation algorithm results in assisting oncologists to make easy decisions for diagnostic purposes. This paper presents an automatic classification of MR brain images in normal and malignant conditions. The feature extraction is done with gray-level co-occurrence matrix, and we proposed a feature reduction technique based on statistical test which is preceded by principal component analysis (PCA). The main focus of the work is to establish the statistical significance of the features obtained after PCA, thereby selecting significant feature values for subsequent classification. For that, a $t$-test is performed which yielded a $p$-value of 0.05. Finally, a comparative study using $k$-nearest neighbor (kNN), support vector machine and artificial neural network (ANN)-based supervised classifiers is performed. In this work, we could achieve reasonably good sensitivity, specificity and accuracy for all the classifiers. The ANN classifier gives better performance with sensitivity of 97.33%, specificity of 97.42% and accuracy of 98.66% on the whole brain atlas database. The experimental results obtained are comparable to the other recent state-of-the-art.