No abstract

Gastrointestinal (GI) disease is one of the most common diseases and primarily examined by GI endoscopy. Recently, deep learning (DL), in particular convolutional neural networks (CNNs) have made achievements in GI endoscopy image analysis. This review focuses on the applications of DL methods in the analysis of GI images. We summarized and compared the latest published literature related to the common clinical GI diseases and covers the key applications of DL in GI image detection, classification, segmentation, recognition, location, and other tasks. At the end, we give a discussion on the challenges and the research directions of GI image analysis based on DL in the future.

Gastrointestinal (GI) diseases constitute a leading problem in the human digestive system. Consequently, several studies have explored automatic classification of GI diseases as a means of minimizing the burden on clinicians and improving patient outcomes, for both diagnostic and treatment purposes. The challenge in using deep learning-based (DL) approaches, specifically a convolutional neural network (CNN), is that spatial information is not fully utilized due to the inherent mechanism of CNNs. This paper proposes the application of spatial factors in improving classification performance. Specifically, we propose a deep CNN-based spatial attention mechanism for the classification of GI diseases, implemented with encoder–decoder layers. To overcome the data imbalance problem, we adapt data-augmentation techniques. A total of 12,147 multi-sited, multi-diseased GI images, drawn from publicly available and private sources, were used to validate the proposed approach. Furthermore, a five-fold cross-validation approach was adopted to minimize inconsistencies in intra- and inter-class variability and to ensure that results were robustly assessed. Our results, compared with other state-of-the-art models in terms of mean accuracy (ResNet50 = 90.28, GoogLeNet = 91.38, DenseNets = 91.60, and baseline = 92.84), demonstrated better outcomes (Precision = 92.8, Recall = 92.7, F1-score = 92.8, and Accuracy = 93.19). We also implemented t-distributed stochastic neighbor embedding (t–SNE) and confusion matrix analysis techniques for better visualization and performance validation. Overall, the results showed that the attention mechanism improved the automatic classification of multi-sited GI disease images. We validated clinical tests based on the proposed method by overcoming previous limitations, with the goal of improving automatic classification accuracy in future work.

Studies had found that bacterial genes are preferentially located on the leading strands. Subsequently, the preferences of essential genes and highly expressed genes were compared by classifying all genes into four groups, which showed that the former has an exclusive influence on orientation. However, only some functional classes of essential genes have this orientation bias. Nevertheless, previous studies only performed comparative analyzes by differentiating the orientation bias extent of two types of genes. Thus, it is unclear whether the influence of essentiality on strand bias works continuously. Herein, we found a significant correlation between essentiality and orientation bias extent in 19 of 21 analyzed bacterial genomes, based on quantitative measurement of gene essentiality (or fitness). The correlation coefficient was much higher than that derived from binary essentiality measures (essential or non-essential). This suggested that genes with relatively lower essentiality, i.e., conditionally essential genes, also have some orientation bias, although it is weaker than that of absolutely essential genes. The results demonstrated the continuous influence of essentiality on orientation bias and provided details on this visible structural feature of bacterial genomes. It also proved that Geptop and IFIM could serve as useful resources of bacterial gene essentiality, particularly for quantitative analysis.

No abstract