Siyuan Lu

This study overviews the application of artificial intelligence (AI) in the prediction of biomarkers for non-small cell lung cancer (NSCLC). Different from recent reviews on AI-driven radiogenomics that adopt clinical or imaging-based classification, this work innovatively takes AI model architecture as the core classification framework. By analyzing multimodal data (mainly CT) and modeling paradigms (radiomics, deep learning and fusion models), it reveals the current research status centered on EGFR mutation prediction. In terms of technological evolution, prediction models are evolving from traditional radiomics to end-to-end learning, Transformer architecture and multimodal perception-driven deep learning paradigms. This study deeply analyzes the performance differences among different models, fills the application gaps in existing reviews, and provides an important benchmark reference for the clinical optimization of AI-assisted tools and the promotion of standardized precise diagnosis and treatment of NSCLC.

Blood cells play an important role in the metabolism of the human body, and the status of blood cells can be used for clinical diagnoses, such as the ratio of different blood cells. Therefore, blood cell classification is a primary task, which requires much time for manual analysis. The recent advances in computer vision can be beneficial to free doctors from tedious tasks. In this paper, a novel automated blood cell classification model based on the shifted window vision transformer (SW-ViT) is proposed. The SW-ViT architecture is firstly pre-trained on the ImageNet dataset and fine-tuned on the blood cell images for classification. Two transfer strategies are employed to generate better classification results. One is to fine-tune the entire SW-ViT, and the other is to only fine-tune the linear output layer of the SW-ViT while all the other parameters are frozen. A public dataset named BCCD_Dataset (Blood Cell Count and Detection) is utilized in the experiments. The results show that the SW-ViT outperforms several state-of-the-art methods in terms of classification accuracy. The proposed SW-ViT can be applied in daily clinical diagnosis.

Background: Blood is responsible for delivering nutrients to various organs, which store important health information about the human body. Therefore, the diagnosis of blood can indirectly help doctors judge a person’s physical state. Recently, researchers have applied deep learning (DL) to the automatic analysis of blood cells. However, there are still some deficiencies in these models. Methods: To cope with these issues, we propose a novel network for the multi-classification of blood cells, which is called DLBCNet. A new specifical model for blood cells (BCGAN) is designed to generate synthetic images. The pre-trained ResNet50 is implemented as the backbone model, which serves as the feature extractor. The extracted features are fed to the proposed ETRN to improve the multi-classification performance of blood cells. Results: The average accuracy, average sensitivity, average precision, average specificity, and average f1-score of the proposed model are 95.05%, 93.25%, 97.75%, 93.72%, and 95.38%, accordingly. Conclusions: The performance of the proposed model surpasses other state-of-the-art methods in reported classification results.

COVID-19 is a contagious infection that has severe effects on the global economy and our daily life. Accurate diagnosis of COVID-19 is of importance for consultants, patients, and radiologists. In this study, we use the deep learning network AlexNet as the backbone, and enhance it with the following two aspects: 1) adding batch normalization to help accelerate the training, reducing the internal covariance shift; 2) replacing the fully connected layer in AlexNet with three classifiers: SNN, ELM, and RVFL. Therefore, we have three novel models from the deep COVID network (DC-Net) framework, which are named DC-Net-S, DC-Net-E, and DC-Net-R, respectively. After comparison, we find the proposed DC-Net-R achieves an average accuracy of 90.91% on a private dataset (available upon email request) comprising of 296 images while the specificity reaches 96.13%, and has the best performance among all three proposed classifiers. In addition, we show that our DC-Net-R also performs much better than other existing algorithms in the literature.

Accurate and timely diagnosis of COVID-19 is indispensable to control its spread. This study proposes a novel explainable COVID-19 diagnosis system called CGENet based on graph embedding and an extreme learning machine for chest CT images. We put forward an optimal backbone selection algorithm to select the best backbone for the CGENet based on transfer learning. Then, we introduced graph theory into the ResNet-18 based on the k-nearest neighbors. Finally, an extreme learning machine was trained as the classifier of the CGENet. The proposed CGENet was evaluated on a large publicly-available COVID-19 dataset and produced an average accuracy of 97.78% based on 5-fold cross-validation. In addition, we utilized the Grad-CAM maps to present a visual explanation of the CGENet based on COVID-19 samples. In all, the proposed CGENet can be an effective and efficient tool to assist COVID-19 diagnosis.

Extreme learning machine (ELM) is a training algorithm for single hidden layer feedforward neural network (SLFN), which converges much faster than traditional methods and yields promising performance. In this paper, we hope to present a comprehensive review on ELM. Firstly, we will focus on the theoretical analysis including universal approximation theory and generalization. Then, the various improvements are listed, which help ELM works better in terms of stability, efficiency, and accuracy. Because of its outstanding performance, ELM has been successfully applied in many real-time learning tasks for classification, clustering, and regression. Besides, we report the applications of ELM in medical imaging: MRI, CT, and mammogram. The controversies of ELM were also discussed in this paper. We aim to report these advances and find some future perspectives.

It is significant to classify teeth categories in dental treatment. A novel teeth classification method was proposed in this paper, which combined fractional Fourier entropy and feedforward neural network. Firstly, fractional Fourier transform was performed on the teeth CT images and the obtained spectrums were used to extract entropies as the features. Then, a feedforward neural network was employed for automatic classification. To train the parameters in the network, improved hybrid genetic algorithm was leveraged. Experiment results suggested that our method achieved state-of-the-art performance.