Breast cancer detection from biopsy images using nucleus guided transfer learning and belief based fusion.

BACKGROUND AND OBJECTIVE: Breast cancer is a frequently diagnosed cancer in women, contributing to significant mortality rates. Death rates are relatively higher in developing nations due to the shortage of early detection amenities and constraints on access to technical advances combating this disease. The only way to diagnose cancer with certainty is through biopsy performed by pathologists. Computer-aided diagnostic algorithms can assist pathologists in being more productive, objective and consistent in the diagnostic process. The focus of this work is to develop a reliable automated breast cancer diagnosis method which can operate in the prevailing clinical environment. METHODS: Nuclei overlap and complex structural organisation of the breast tissue in biopsy images make nuclei segmentation, feature extraction and classification challenging. In this work, a nucleus guided transfer learning (NucTraL) methodology is proposed as a simple and affordable breast tumor classification algorithm. The image feature is represented by fusion of local nuclei features that are extracted using convolutional neural network (CNN) models pretrained on the ImageNet database. The nucleus patch extraction strategy used in this work avoids fine segmentation of the nuclei boundary but provides features with good discriminative power for classification. Classification of the fused features into benign and malignant classes is performed using a support vector machine (SVM) classifier. A belief theory based classifier fusion (BCF) strategy is then employed to combine the outputs arising from the different CNN-SVM combinations to improve accuracy further. RESULTS: Evaluation of results is achieved by executing 100 random trials with 70%-30% train to test division on the publicly available BreaKHis dataset. The proposed framework achieved average accuracy of 96.91%, sensitivity of 97.24% and specificity of 96.18%. CONCLUSION: It is found that the proposed NucTraL+BCF framework outperforms several recent approaches and achieves results comparable to the state-of-the-art methods even without using high computational power. This qualitative framework based on transfer learning can contribute significantly for developing cost effective and low complexity CAD system for breast cancer diagnosis from histopathological images.

Chaotic analysis of the electroretinographic signal for diagnosis.

Electroretinogram (ERG) is a time-varying potential which arises from different layers of retina. To be specific, all the physiological signals may contain some useful information which is not visible to our naked eye. However this subtle information is difficult to monitor directly. Therefore the ERG signal features which are extracted and analyzed using computers are highly useful for diagnosis. This work discusses the chaotic aspect of the ERG signal for the controls, congenital stationary night blindness (CSNB), and cone-rod dystrophy (CRD) classes. In this work, nonlinear parameters like Hurst exponent (HE), the largest Lyapunov exponent (LLE), Higuchi's fractal dimension (HFD), and approximate entropy (ApEn) are analyzed for the three different classes. It is found that the measures like HE dimension and ApEn are higher for controls as compared to the other two classes. But LLE shows no distinguishable variation for the three cases. We have also analyzed the recurrence plots and phase-space plots which shows a drastic variation among the three groups. The results obtained show that the ERG signal is highly complex for the control groups and less complex for the abnormal classes with P value less than 0.05.

Analysis of myocardial infarction using discrete wavelet transform.

Myocardial infarction (MI), is commonly known as a heart attack, occurs when the blood supply to the portion of the heart is blocked causing some heart cells to die. This information is depicted in the elevated ST wave, increased Q wave amplitude and inverted T wave of the electrocardiogram (ECG) signal. ECG signals are prone to noise during acquisition due to electrode movement, muscle tremor, power line interference and baseline wander. Hence, it becomes difficult to decipher the information about the cardiac state from the morphological changes in the ECG signal. These signals can be analyzed using different signal processing techniques. In this work, we have used multiresolution properties of wavelet transformation because it is suitable tool for interpretation of subtle changes in the ECG signal. We have analyzed the normal and MI ECG signals. ECG signal is decomposed into various resolution levels using the discrete wavelet transform (DWT) method. The entropy in the wavelet domain is computed and the energy-entropy characteristics are compared for 2282 normal and 718 MI beats. Our proposed method is able to detect the normal and MI ECG beat with more than 95% accuracy.