Cancer Research
Current Projects
Preprocessing of Breast Cancer Images to Create Datasets for Deep-CNN
Breast cancer is the most diagnosed cancer in Australia with crude incidence rates increasing drastically from 62.8 at ages 35-39 to 271.4 at ages 50-54 (cases per 100,000 women). Various researchers have proposed methods and tools based on Machine Learning and Convolutional Neural Networks for assessing mammographic images, but these methods have produced detection and interpretation errors resulting in false-positive and false-negative cases when used in the real world. We believe that this problem can potentially be resolved by implementing effective image pre-processing techniques to create training data for Deep-CNN. Therefore, the main aim of this research is to propose effective image pre-processing methods to create datasets that can save computational time for the neural network and improve accuracy and classification rates. To do so, this research proposes methods for background removal, pectoral muscle removal, adding noise to the images, and image enhancements. Adding noise without affecting the quality of details in the images makes the input images for the neural network more representative, which may improve the performance of the neural network model when used in the real world. The proposed method for background removal is the “Rolling Ball Algorithm” and “Huang’s Fuzzy Thresholding”, which succeed in removing background from 100% of the images. For pectoral muscle removal “Canny Edge Detection” and “Hough’s Line Transform” are used, which removed muscle from 99.06% of the images. “Invert”, “CTI_RAS” and “ISOCONTOUR” lookup tables (LUTs) were used for image enhancements to outline the ROIs and regions within the ROI
Classification of Enhanced Mammogram Images using D-CNN, C-ReLU and AM-SoftMax Functions
Accurate classification of mammogram images can lead to early detection and appropriate treatment of breast cancer. However, distinguishing benign and malignant masses and calcification in these images is challenging due to the large intraclass and small interclass variations. This research proposes a D-CNN model to classify the images of the CBIS-DDSM dataset into four classes (benign calcification, benign mass, malignant calcification, and malignant mass). After processing and enhancing the original images to create training, validation, and testing datasets, we have used Concatenated-Rectified Linear Unit (CReLU) and Additive Margin SoftMax (AM-SoftMax) functions to construct our D-CNN model. The proposed D-CNN model achieved a testing accuracy of 93.99% using the processed images. It has better accuracy and lower network depth than other current state-of-the-art CNN architectures
Classification of Cancers using Nuclei segmentation
Nuclei segmentation is a process of detecting and delineating each nucleus in microscopy images. It extracts the interpretable features of diagnostic and prognostic cancer indicators and regarded as the crucial step for précising the medicine. Instead of manual nuclei segmentation (H&E stained based biopsy), automated nuclei segmentation (semantic and instance) methods are gaining popularity. However, the automatic process still remains a challenging task in terms of robustness. This project intents to develop a deep neural network model that can perform tissue-wise and nuclei category-wise classification, detection, and semantic and instance segmentations.
A Performance based Study on Deep Learning Algorithms in the Effective Prediction of Breast Cancer
Breast Cancer is one of the leading causes of death worldwide. Early detection is very important in increasing survival rates. Intensive research is therefore done to improve early detection of such cancers through the use of available technology. This includes various image processing techniques and general machine learning. However, the reported accuracy for many of these studies was often not at the desirable level. Deep Learning based techniques are a promising approach for the early detection of Breast Cancer. We have therefore done a comparative analysis of seven Deep Learning techniques applied to the Wisconsin Breast Cancer (Diagnostic) Dataset.
