Skip to Main Content CMEIASBannerImages

CMEIAS Morphotype Classification

Step-by-Step Demo

The CMEIAS Morphotype Classifier uses a series of pattern recognition algorithms operating in 14-dimensional space and optimized by us to classify 11 major microbial morphotypes representing 98% of the genera described in the 9th Edition of Bergey's Manual of Determinative Bacteriology. Extensive testing using large ground truth data sets indicate that CMEIAS performs this morphotype classification with an overall accuracy of 97% on properly edited images (Liu et al. 2001).

To perform a morphotype classification using CMEIAS-IT v. 1.28, the user first finds the foreground objects of interest in the image by using a grayscale brightness thresholding procedure, then conducts an Object Analysis to extract 14 shape measurement attributes from each microbe present, and finally uses these Object Analysis data to perform object classification that automatically assigns the appropriate morphotype to each microbe found.

The CMEIAS data output is a results window that reports on the richness of different morphotypes found and the distribution of abundance among each of them. These data can be used to compute various indices of morphological diversity and numerical abundance in microbial community analysis. An interactive edit feature is included to address the main sources of error in automatic morphotype classification, enabling the user to inspect the morphotype assigned to each microbe in the image based on visual recognition of its distinctive pseudocolor, reassign it to another morphotype class if necessary, and add up to five other rare morphotypes to the supervised classification scheme.

Comprehensive descriptions of the morphotype classifier in CMEIAS-IT v. 1.28, including its rationale for development, algorithms of size and shape measurements, rules of pattern recognition for each morphotype classification, results of thorough accuracy testing, sources of error and limitations, and an example of how CMEIAS can augment the polyphasic analysis of growing, complex microbial communities are presented in Liu et al. 2001 and Dazzo and Niccum 2015. A user manual, interactive training tutorial, and macros to run object analysis and object classification on your images are described in the CMEIAS user support section and also provided in the CMEIAS-IT v. 1.28 download. The CMEIAS 1.28 interactive training tutorial includes a section of image analysis using this object classification routine.



Step-by-Step Demo of the CMEIAS Morphotype Classifier:

In this example, the major steps to analyze the morphological diversity of two community images using CMEIAS-IT v. 1.28 are presented. A more detailed description of these image analysis steps is available in the user manual and the interactive training tutorial files provided in the CMEIAS-IT v. 1.28 download.

1.) A high quality 24-bit RGB or 8-bit grayscale digital image is first edited in an image editing program such as CMEIAS Color Segmentation or Adobe Photoshop so that all foreground objects of interest have a brightness range outside that of background, and then converted into a binary image using the brightness thresholding feature in CMEIAS-IT v. 1.28. In this example, two binary composite images were made, each contain 170 microbes representing the distribution of morphotype diversity and abundance in different (A and B) anaerobic bioreactor communities.

Communities A and B

top/menuup arrow

2.) Set preferences and measurement features in UTHSCSA ImageTool to begin CMEIAS morphotype classification.

measurement features

top/menuup arrow

3.) CMEIAS/ImageTool then finds the foreground objects (microbes) in the binary images using a thresholding procedure, annotates a numbered label to each microbe found, and reports the total object count in the image.

Object Annotation Count

top/menuup arrow

4.) Next, CMEIAS/ImageTool analyzes each object in the image and reports the values for each shape measurement feature selected in a Results window worksheet.

Objects Analysis Results Window

top/menuup arrow

5.) CMEIAS then uses these extracted analysis data to perform object classification (identify morphotypes), tabulates the classification count data in the Results window, and creates a new classification image in which each microbe is distinctively pseudocolored according to its assigned morphotype.

Community A Classification

top/menuup arrow

6.) CMEIAS then allows you to inspect the classification assignments for each object in the pseudocolored image, and if necessary (errors occur at ~3% rate), to manually edit individual morphotype classifications using the Reassign Category Label interactive edit feature...

Edit Classifications

as illustrated below using the pseudocolored classification result images of community B.

Community B Before and After Classified

To illustrate the Reassign Classification Label feature, the red coccobacillus (white arrow, center of left image) is reclassified as a regular straight rod morphotype (blue, right image) and the corresponding classification data are automatically updated.

top/menuup arrow

7.) Community morphotype analysis data can then be copied to the clipboard and exported to Windows-compatible spreadsheet programs where they can be graphed and analyzed statistically.

output 1   output 3
Rank Abundance graph

output 4

The Bottom Line

CMEIAS analysis indicates that communities A and B have 48.82% proportional similarity in morphological diversity, with Community B being 2.2-fold higher in morphotype richness and diversity indices and 1.6-fold higher in distribution of morphotype evenness.