Development of an ensemble CNN model with explainable AI for the classification of gastrointestinal cancer.

The implementation of AI assisted cancer detection systems in clinical environments has faced numerous hurdles, mainly because of the restricted explainability of their elemental mechanisms, even though such detection systems have proven to be highly effective. Medical practitioners are skeptical about adopting AI assisted diagnoses as due to the latter's inability to be transparent about decision making processes. In this respect, explainable artificial intelligence (XAI) has emerged to provide explanations for model predictions, thereby overcoming the computational black box problem associated with AI systems. In this particular research, the focal point has been the exploration of the Shapley additive explanations (SHAP) and local interpretable model-agnostic explanations (LIME) approaches which enable model prediction explanations. This study used an ensemble model consisting of three convolutional neural networks(CNN): InceptionV3, InceptionResNetV2 and VGG16, which was based on averaging techniques and by combining their respective predictions. These models were trained on the Kvasir dataset, which consists of pathological findings related to gastrointestinal cancer. An accuracy of 96.89% and F1-scores of 96.877% were attained by our ensemble model. Following the training of the ensemble model, we employed SHAP and LIME to analyze images from the three classes, aiming to provide explanations regarding the deterministic features influencing the model's predictions. The results obtained from this analysis demonstrated a positive and encouraging advancement in the exploration of XAI approaches, specifically in the context of gastrointestinal cancer detection within the healthcare domain.

Identification of Crown of Thorns Starfish (COTS) using Convolutional Neural Network (CNN) and attention model.

Coral reefs play important roles in the marine ecosystem, from providing shelter to aquatic lives to being a source of income to others. However, they are in danger from outbreaks of species like the Crown of Thorns Starfish (COTS) and the widespread coral bleaching from rising sea temperatures. The identification of COTS for detecting outbreaks is a challenging task and is often done through snorkelling and diving activities with limited range, where strong currents result in poor image capture, damage of capturing equipment, and are of high risks. This paper proposes a novel approach for the automatic detection of COTS based Convolutional Neural Network (CNN) with an enhanced attention module. Different pre-trained CNN models, namely, VGG19 and MobileNetV2 have been applied to our dataset with the aim of detecting and classifying COTS using transfer learning. The architecture of the pre-trained models was optimised using ADAM optimisers and an accuracy of 87.1% was achieved for VGG19 and 80.2% for the MobileNetV2. The attention model was developed and added to the CNN to determine which features in the starfish were influencing the classification. The enhanced model attained an accuracy of 92.6% while explaining the causal features in COTS. The mean average precision of the enhanced VGG-19 with the addition of the attention model was 95% showing an increase of 2% compared to only the enhanced VGG-19 model.

Multi-Class Skin Problem Classification Using Deep Generative Adversarial Network (DGAN).

The lack of annotated datasets makes the automatic detection of skin problems very difficult, which is also the case for most other medical applications. The outstanding results achieved by deep learning techniques in developing such applications have improved the diagnostic accuracy. Nevertheless, the performance of these models is heavily dependent on the volume of labelled data used for training, which is unfortunately not available. To address this problem, traditional data augmentation is usually adopted. Recently, the emergence of a generative adversarial network (GAN) seems a more plausible solution, where synthetic images are generated. In this work, we have developed a deep generative adversarial network (DGAN) multi-class classifier, which can generate skin problem images by learning the true data distribution from the available images. Unlike the usual two-class classifier, we have developed a multi-class solution, and to address the class-imbalanced dataset, we have taken images from different datasets available online. One main challenge faced during our development is mainly to improve the stability of the DGAN model during the training phase. To analyse the performance of GAN, we have developed two CNN models in parallel based on the architecture of ResNet50 and VGG16 by augmenting the training datasets using the traditional rotation, flipping, and scaling methods. We have used both labelled and unlabelled data for testing to test the models. DGAN has outperformed the conventional data augmentation by achieving a performance of 91.1% for the unlabelled dataset and 92.3% for the labelled dataset. On the contrary, CNN models with data augmentation have achieved a performance of up to 70.8% for the unlabelled dataset. The outcome of our DGAN confirms the ability of the model to learn from unlabelled datasets and yet produce a good diagnosis result.

Multi- class classification of breast cancer abnormalities using Deep Convolutional Neural Network (CNN).

The real cause of breast cancer is very challenging to determine and therefore early detection of the disease is necessary for reducing the death rate due to risks of breast cancer. Early detection of cancer boosts increasing the survival chance up to 8%. Primarily, breast images emanating from mammograms, X-Rays or MRI are analyzed by radiologists to detect abnormalities. However, even experienced radiologists face problems in identifying features like micro-calcifications, lumps and masses, leading to high false positive and high false negative. Recent advancement in image processing and deep learning create some hopes in devising more enhanced applications that can be used for the early detection of breast cancer. In this work, we have developed a Deep Convolutional Neural Network (CNN) to segment and classify the various types of breast abnormalities, such as calcifications, masses, asymmetry and carcinomas, unlike existing research work, which mainly classified the cancer into benign and malignant, leading to improved disease management. Firstly, a transfer learning was carried out on our dataset using the pre-trained model ResNet50. Along similar lines, we have developed an enhanced deep learning model, in which learning rate is considered as one of the most important attributes while training the neural network. The learning rate is set adaptively in our proposed model based on changes in error curves during the learning process involved. The proposed deep learning model has achieved a performance of 88% in the classification of these four types of breast cancer abnormalities such as, masses, calcifications, carcinomas and asymmetry mammograms.

Studying the trend of the novel coronavirus series in Mauritius and its implications.

Mauritius stands as one of the few countries in the world to have controlled the current pandemic, the novel coronavirus 2019 (COVID-19) to a significant extent in a relatively short lapse of time. Owing to uncertainties and crisis amid the pandemic, as an emergency announcement, the World Health Organization (WHO) solicits the help of health authorities, especially, researchers to conduct in-depth research on the evolution and treatment of COVID-19. This paper proposes an integer-valued time series model to analyze the series of COVID-19 cases in Mauritius wherein the corresponding innovation term accommodates for covariate specification. In this set-up, sanitary curfew followed by sanitization and sensitization campaigns, time factor and safe shopping guidelines have been tested as the most significant variables, unlike climatic conditions. The over-dispersion estimates and the serial auto-correlation parameter are also statistically significant. This study also confirms the presence of some unobservable effects like the pathological genesis of the novel coronavirus and environmental factors which contribute to rapid propagation of the zoonotic virus in the community. Based on the proposed COM-Poisson mixture models, we could predict the number of COVID-19 cases in Mauritius. The forecasting results provide satisfactory mean squared errors. Such findings will subsequently encourage the policymakers to implement strict precautionary measures in terms of constant upgrading of the current health care and wellness system and re-enforcement of sanitary obligations.