Program

Breast cancer is the most common form of cancer amongst women. It is a worldwide problem and a growing one. The data in this presentation will come from breast images and the problem is to detect breast cancer. First, this talk will introduce the problem and highlight some technical challenges. Second, some attempts at breast cancer detection from mammographic images will be outlined and some results will be presented. This presentation will introduce a number of topics from artificial intelligence like feature selection, feature generation, and classification. These will be used in conjunction with image processing to address computer-aided breast cancer detection successfully.

 Malaria is the leading cause of morbidity and mortality in tropical and subtropical countries. The diagnosis is usually done manually by conventional light microscopy which has occasionally proved inefficient since it is laborious, time consuming and subjective. All diagnosis techniques which yield better results are quite expensive and hence inaccessible to the developing countries where this disease is more endemic. To support manual methods, we developed a system which can automatically detect the plasmodium parasites from thin blood smear images. We used a dataset provided by the Centers for Disease Control and Prevention (CDC). Although automating the process is a good solution, the available techniques are unable to evaluate the same cases at which the morphology of red blood cells (RBCs) is affected by diseases such as anemia and hemoglobinopathies. In this work, we proposed an accurate, rapid and affordable system for malaria diagnosis. The proposed method applies image preprocessing with bilateral filters, followed by the contrast-limited adaptive histogram equalization (CLAHE), and color-based segmentation of parasites through an adaptive Gaussian mixture model (GMM). We investigated the detection performance with color features from the CIElab and HSI color spaces in addition to the RGB colour model. Then, occluded RBCs were separated using the distance transform, and local maxima of the distance map where detected as the centers of the RBC cells. Overlapping cells can then be separated using the shortest-path methods. Further, the classification and counting of infected and non-infected RBCs have been made. The experimental results demonstrated that the developed method had a good performance (98.98% accuracy, 95.16% sensitivity and 99.58% specificity) with respect to the ground truth data. Eventually, this work can be considered as a low-cost solution for the malaria quantification in lower income countries and massive examinations.

Genetic variation dramatically affects our lives. It affects what we look like, our personalities and preferences. It also important for adaptation. Common genetic variation also affects the metabolism, plasma availability, or clinical response to drugs. More crucially, genetic variation affects susceptibility and severity to disease. Scientists are just beginning to understand how these genes interact with each other and with environmental factors in ways that impact on health. The study of genetic variation has been used to model human migration, understand cause of human diseases, and to predict disease outcomes.
In this talk, I am going to introduce some key concepts in the field of human genetic variation including the types of variants, variant effects and genetic association studies. Additionally, I will give you a hint about my research about “ Haplotype blocks and SNPs”.

Monitoring the evolution pattern of Mammals lung adenocarcinoma genes

Electronic health records (EHR) are represented at any patient reports or image-based datasets generated by histopathology, pathology and radiography. Nowadays, there is an urgent need to extract useful clinical information out of EHR that may help in complete understanding of diseases. However, the enormous flow of data may have its impact upon the feasibility of useful information extraction as it consumes time and effort. Machine learning models can handle these large data sets and extract informative features that can be used in consequent diagnostic approaches. In this review, I illustrate different means of machine learning models and its applications in the diagnosis of cancer and diabetes.

The fourth industrial revolution is affecting all aspects of our lives and healthcare is one of the top industries to be affected. Digital Health is here to stay and we have to prepare for it. In building the hospitals of the future we will look at how they are forecast to change, how the workforce will evolve and what technologies are shaping this revolution. We’ll focus on the multidisciplinary input needed to make it happen and how we can make a change that starts from home.