Segmentation, Splitting, and Classification of Overlapping Bacteria in Microscope Images for Automatic Bacterial Vaginosis Diagnosis

Abstract

Quantitative analysis of bacterial morphotypes in the microscope images plays a vital role in diagnosis of bacterial vaginosis (BV) based on the Nugent score criterion. However, there are two main challenges for this task: 1) It is quite difficult to identify the bacterial regions due to various appearance, faint boundaries, heterogeneous shapes, low contrast with the background, and small bacteria sizes with regards to the image. 2) There are numerous bacteria overlapping each other, which hinder us to conduct accurate analysis on individual bacterium. To overcome these challenges, we propose an automatic method in this paper to diagnose BV by quantitative analysis of bacterial morphotypes, which consists of a three-step approach, i.e., bacteria regions segmentation, overlapping bacteria splitting, and bacterial morphotypes classification. Specifically, we first segment the bacteria regions via saliency cut, which simultaneously evaluates the global contrast and spatial weighted coherence. And then Markov random field model is applied for high-quality unsupervised segmentation of small object. We then decompose overlapping bacteria clumps into markers, and associate a pixel with markers to identify evidence for eventual individual bacterium splitting. Next, we extract morphotype features from each bacterium to learn the descriptors and to characterize the types of bacteria using an Adaptive Boosting machine learning framework. Finally, BV diagnosis is implemented based on the Nugent score criterion. Experiments demonstrate that our proposed method achieves high accuracy and efficiency in computation for BV diagnosis.

Morphologic Classification and Automatic Diagnosis of Bacterial Vaginosis by Deep Neural Networks

“…To better understand the performance of the best model on BV classification, the three classification results (Confusion Matrix) of the best point of AUC curve for the 1/4 NugentNet was plotted against the labeled results of the test set. The best point obtained 89.3% accuracy on the three-category classification, which was 5.6% higher than the microscopists and comparable to the top experts' performance in China [17]. The results showed only 3.8% (20/531) BV samples were predicted as false negative (normal vaginal flora) and 0.1% (6/4120) normal vaginal flora samples were predicted as false positive (BV).…”

“…These morphotypes were thought to represent Lactobacillus spp., Gardnerella vaginalis and Mobiluncus spp., respectively. Nugent scoring had since then become the 'gold standard' for laboratory diagnosis of BV [7,8,17]. In Nugent scale, scores of 0-3 were considered to have normal vaginal flora (Lactobacillus dominant); scores of 4-6 were labeled as altered vaginal flora (mixed morphotypes); and scores of 7-10 were indicative of BV (absence of lactobacilli and predominance of the other 2 morphotypes).…”

“…Alternative diagnostic methods such as molecular diagnostic assays, enzymatic assays, and chromogenic point-of-care test (POCT) were compared to Nugent criteria [18]. However, the determination of a Nugent score by a microbiologist could be easily influenced by individual skill and was time consuming [17,18]. In addition, the number of experienced microbiologists or technologist performing the microscopic work was in shortage among different countries and districts [19].…”

“…Recently, a traditional image processing method was developed for automatic bacterial vaginosis diagnosis. However, the sensitivity (58.3%) and specificity (79.1%) was relatively poor in comparison to expert due to its limitation in the image processing algorithm [17]. Deep learning method especially convolutional neural network (CNN) models had demonstrated excellent performance on computer vision tasks including image classification, image semantic segmentation and image object detection.…”

An effective and accurate identification system of Mycobacterium tuberculosis using convolution neural networks

“…Many conventional systems, such as those developed by MoradiAmin, Memari, Samadzadehaghdam, Kermani, & Talebi (2016), Khutlang, Krishnan, Whitelaw, &Douglas (2010) andCostaFilho et al (2012), followed the detection-classification stage structure and have achieved satisfactory results. Song et al (2017) and Momenzadeh, Vard, Talebi, Mehri Dehnavi, & Rabbani (2018) proposed the computer-aided diagnostic systems for microscopic images using segmentation and classification methods. Zhai, Liu, Zhou, & Liu (2010) and Shah, Mishra, Sarkar, & Sudarshan (2016) proposed a fully automated M. tuberculosis identification system, consisting of image capturing, microscopy system setting, and identification methods.…”