Large-scale characterisation of the pregnancy vaginal microbiome and sialidase activity in a low-risk Chinese population

Vaginal microbiota-host interactions are linked to preterm birth (PTB), which continues to be the primary cause of global childhood mortality. Due to population size, the majority of PTB occurs in Asia, yet there have been few studies of the pregnancy vaginal microbiota in Asian populations. Here, we characterized the vaginal microbiome of 2689 pregnant Chinese women using metataxonomics and in a subset (n = 819), the relationship between vaginal microbiota composition, sialidase activity and leukocyte presence and pregnancy outcomes. Vaginal microbiota were most frequently dominated by Lactobacillus crispatus or L. iners, with the latter associated with vaginal leukocyte presence. Women with high sialidase activity were enriched for bacterial vaginosis-associated genera including Gardnerella, Atopobium and Prevotella. Vaginal microbiota composition, high sialidase activity and/or leukocyte presence was not associated with PTB risk suggesting underlying differences in the vaginal microbiota and/or host immune responses of Chinese women, possibly accounting for low PTB rates in this population.

Morphological Components Detection for Super-Depth-of-Field Bio-Micrograph Based on Deep Learning

Accompanied with the clinical routine examination demand increase sharply, the efficiency and accuracy are the first priority. However, automatic classification and localization of cells in microscopic images in super depth of Field (SDoF) system remains great challenges. In this paper, we advance an object detection algorithm for cells in the SDoF micrograph based on Retinanet model. Compared with the current mainstream algorithm, the mean average precision (mAP) index is significantly improved. In the experiment of leucorrhea samples and fecal samples, mAP indexes are 83.1% and 88.1%, respectively, with an average increase of 10%. The object detection model proposed in this paper can be applied to feces and leucorrhea detection equipment, and significantly improve the detection efficiency and accuracy.

Fast and accurate automated recognition of the dominant cells from fecal images based on Faster R-CNN

Fecal samples can easily be collected and are representative of a person’s current health state; therefore, the demand for routine fecal examination has increased sharply. However, manual operation may pollute the samples, and low efficiency limits the general examination speed; therefore, automatic analysis is needed. Nevertheless, recognition exhaustion time and accuracy remain major challenges in automatic testing. Here, we introduce a fast and efficient cell-detection algorithm based on the Faster-R-CNN technique: the Resnet-152 convolutional neural network architecture. Additionally, a region proposal network and a network combined with principal component analysis are proposed for cell location and recognition in microscopic images. Our algorithm achieved a mean average precision of 84% and a 723 ms detection time per sample for 40,560 fecal images. Thus, this approach may provide a solid theoretical basis for real-time detection in routine clinical examinations while accelerating the process to satisfy increasing demand.

FecalNet: Automated detection of visible components in human feces using deep learning

PURPOSE

To automate the detection and identification of visible components in feces for early diagnosis of gastrointestinal diseases, we propose FecalNet, a method using multiple deep neural networks.

METHODS

FecalNet uses the ResNet152 residual network to extract and learn the characteristics of visible components in fecal microscopic images, acquire feature maps in combination with the feature pyramid network, apply the full convolutional network to classify and locate the fecal components, and implement the improved focal loss function to reoptimize the classification results. This allowed the complete automation of the detection and identification of the visible components in feces.

RESULTS

We validated this method using a fecal database of 1,122 patients. The results indicated a mean average precision (mAP) of 92.16% and an average recall (AR) of 93.56%. The average precision (AP) and AR of erythrocyte, leukocyte, intestinal mucosal epithelial cells, hookworm eggs, ascarid eggs, and whipworm eggs were 92.82% and 93.38%, 93.99% and 96.11%, 90.71% and 92.41%, 89.95% and 93.88%, 96.90% and 91.21%, and 88.61% and 94.37%, respectively. The average times required by the GPU and the CPU to analyze a fecal microscopic image are approximately 0.14 and 1.02 s, respectively.

CONCLUSION

FecalNet can automate the detection and identification of visible components in feces. It also provides a detection and identification framework for detecting several other types of cells in clinical practice.

A computer-assisted human peripheral blood leukocyte image classification method based on Siamese network

A computer-assisted human peripheral blood leukocyte image classification method based on Siamese network is proposed. Firstly, a Siamese network with two identical convolutional neural network (CNN) sub-networks and a logistic regression for leukocyte five classification is designed, which can learn not only distinguishing features but also a similarity metric. Then for each category of the leukocytes, a typical sample is selected by the hematologist. To train the Siamese network, a leukocyte and a typical sample that belong to the same category make up a genuine pair and the leukocyte with the rest four typical samples respectively make up four impostor pairs. Obviously, the number of the genuine pairs is lesser than that of the impostor pairs. Thus, a data augmentation method suitable for leukocyte is used to enrich the amount of the genuine pairs. By training the Siamese network using the genuine pairs and impostor pairs, the Siamese network can not only shorten the similarity metric between the leukocyte and the same category of the typical sample but also increase the similarity metrics between the leukocyte and the different categories of the typical samples. Experimental results indicate that the proposed method can achieve 98.8% average testing accuracy. Graphical abstract.

Human peripheral blood leukocyte classification method based on convolutional neural network and data augmentation

PURPOSE

Human peripheral blood leukocytes' classification is important for diagnosing blood diseases. Many microscopic leukocyte image automatic detection methods are proposed. In recent years, convolutional neural networks (CNNs) are applied to microscopic leukocyte image automatic classification. But when a CNN is used for microscopic leukocyte image classification, the dataset's scarcity and imbalance will lead to low classification accuracy. To improve classification accuracy, a data augmentation method is proposed, and a resampling method is adopted when using a CNN method.

METHODS

First, a deep CNN model for microscopic leukocyte image classification is designed. Then, a new data augmentation method based on feature concentration is proposed to enrich the dataset and overcome the problem of dataset scarcity. To make the CNN model focus on the leukocyte region, many images are generated by putting a segmented leukocyte into images with different microscopic surroundings using an image processing method. Finally, taking the imbalance of the five kinds of leukocytes in the dataset into consideration, a resampling method is adopted. The resampling method iteratively feeds the leukocyte images with a low proportion to the CNN model within an epoch to ensure that images of each of the five kinds of leukocytes are represented in relatively equal numbers in each batch.

RESULTS

The experimental results demonstrate that the proposed classification method can achieve 97.6% average testing accuracy. Classification precision for the five kinds of leukocytes is above 93.4%, while sensitivity is above 92.5%. Both the proposed data augmentation and the resampling methods improve classification accuracy.

CONCLUSIONS

A human peripheral blood leukocyte classification method based on a CNN and data augmentation is proposed. The problem of dataset scarcity is solved by the proposed data augmentation method, and the dataset imbalance is solved by a resampling method.

Automatic classification of cells in microscopic fecal images using convolutional neural networks

The analysis of fecal-type components for clinical diagnosis is important. The main examination involves the counting of red blood cells (RBCs), white blood cells (WBCs), and molds under the microscopic. With the development of machine vision, some vision-based detection schemes have been proposed. However, these methods have a single target for detection, with low detection efficiency and low accuracy. We proposed an algorithm to identify the visible image of fecal composition based on intelligent deep learning. The algorithm mainly includes region proposal and candidate recognition. In the process of segmentation, we proposed a morphology extraction algorithm in a complex background. As for the candidate recognition, we proposed a new convolutional neural network (CNN) architecture based on Inception-v3 and principal component analysis (PCA). This method achieves high-average Precision of 90.7%, which is better than the other mainstream CNN models. Finally, the images within the rectangle marks were obtained. The total time for detection of an image was roughly 1200 ms. The algorithm proposed in the present paper can be integrated into an automatic fecal detection system.

Leukocyte recognition in human fecal samples using texture features

Unlike urine or blood samples with a single background, human fecal samples contain large amounts of food debris, amorphous particles, and undigested plant cells. It is difficult to segment such impurities when mixed with leukocytes. Cell degradation results in ambiguous nuclei, incompleteness of the cell membrane, and a changeable cell morphology, which are difficult to recognize. Aiming at the segmentation problem, a threshold segmentation method combining an inscribed circle and circumscribed circle is proposed to effectively remove the adhesion impurities with a segmentation accuracy reaching 97.6%. For the identification problem, five texture features (i.e., LBP-uniform, Gabor, HOG, GLCM, and Haar) were extracted and classified using four kinds of classifiers (support vector machine (SVM), artificial neural network, AdaBoost, and random forest). The experimental results show that using a histogram of oriented gradient features with an SVM classifier can achieve precision of 88.46% and recall of 88.72%.