Details

Name
Mouse Fundus Lesion Analysis
IQbot ID
IQBOT-0016
Version
1.0.0
Research Area(s)
Angiogenesis, Diabetes, Drug Toxicity & Metabolism, Genetics, & Ophthalmology
Processing Time Expectation
Less than 2 Minutes per Image
Acquisition Hardware
Funduscope or Ophthalmoscope

 

Name
Mouse Fundus Lesion Analysis
IQbot ID
IQBOT-0016
Version
1.0.0
Research Area(s)
Angiogenesis, Diabetes, Drug Toxicity & Metabolism, Genetics, & Ophthalmology
Processing Time Expectation
Less than 2 Minutes per Image
Acquisition Hardware
Funduscope or Ophthalmoscope

Output Parameters

Total Lesion Count
The number of lesions in the entire field-of-view.
Mean Lesion Area (pixels)
The average area of the segmented lesion set.
Standard Deviation of Lesion Area (pixels)
The standard deviation of all segmented lesion areas.
Minimum Lesion Area (pixels)
The area of the smallest segmented lesion in the field-of-view.
Maximum Lesion Area (pixels)
The area of the largest segmented lesion in the field-of-view.
Total Lesion Area (pixels)
The sum of all lesion areas in the field-of-view.
Lesion Area (pixels)
The area of a given individual lesion (listed by row).

 

This section will identify any output parameters that are only output by the IQbot in its advanced configuration.

 

Total Lesion Count
The number of lesions in the entire field-of-view.
Mean Lesion Area (pixels)
The average area of the segmented lesion set.
Standard Deviation of Lesion Area (pixels)
The standard deviation of all segmented lesion areas.
Minimum Lesion Area (pixels)
The area of the smallest segmented lesion in the field-of-view.
Maximum Lesion Area (pixels)
The area of the largest segmented lesion in the field-of-view.
Total Lesion Area (pixels)
The sum of all lesion areas in the field-of-view.
Nerve Centroid X (pixels)
The x-coordinate of the center of the optic nerve (0 is the left edge of the fundus image).
Nerve Centroid Y (pixels)
The y-coordinate of the center of the optic nerve (0 is the top edge of the fundus image).
Lesion Centroid X (pixels)
The x-coordinate of the center of a given lesion (0 is the left edge of the fundus image).
Lesion Centroid Y (pixels)
The y-coordinate of the center of a given lesion (0 is the top edge of the fundus image).
Lesion Area (pixels)
The area of a given individual lesion (listed by row).

Image Acquisition


This IQbot was developed to analyze images captured with a funduscope or ophthalmoscope. Sections imaged and subsequently submitted for analysis should depict the full range of affected tissue, including the optic nerve and vascualture of the eye.

The Mouse Fundus Lesion Analysis IQbot has been built to accommodate .tif and .tiff file types.

As we reference for IQbots in general here, the Mouse Fundus Lesion Analysis IQbot is only ultimately effective if you upload high quality images for processing. Please ensure that any images you upload follow the below guidelines specific to this IQbot.

High-Resolution:  The quality of the output data you receive is largely dependent on the input image resolution. Please provide high-resolution images to ensure that the IQbot can most effectively differentiate between features.
In-Focus:  Please produce the most in-focus images possible for analysis. Images with high background contrast will result in more accurate and useful data and overlays.
Properly Saturated:  If your submitted images are significantly undersaturated or oversaturated, it is possible that the IQbot will be unable to detect structures, or may group structures together. Please ensure that the images you upload are saturated appropriately.
Clear of Artifacts:  Please take care to avoid creating bubbles, debris, or otherwise inducing artifacts into your image. Such additional features may corrupt the resulting data.
Clear of Accessary Information:  Please do not submit images retaining any annotations, legends, timestamps, or other markings. Every part of the images that you upload will be analyzed and included in the output data.
 
 
 

If you are uncertain whether or not your images meet the quality standards for which this IQbot was validated, or have additional questions relating to image quality, please view our FAQs or contact the ImageQuantify.com support team.

 

     (216) 678-9258           help@imagequantify.com

 

Validation Information


This IQbot has been successfully validated for use, and a copy of its validation certificate accompanies the results of each order it analyzes. In accordance with general ImageQuantify.com validation procedures, this IQbot has been found to provide quantitative information equivalent to that of a manual observer, but with no variability and increased breadth of information.

 
 

The first step of the validation process was to establish the ground truth against which the IQbot’s efficacy would be compared. For the purpose of the Mouse Fundus Lesion Analysis IQbot, ground truth was defined on a manual observer basis. Trained observers were guided through a standardized, semi-automated process to produce output data from a specially selected validation image set. This image set represented the entire range of expected output values (e.g., with low and high lesion density as well as diverse lesion localization). Images were presented for analysis in a randomized order and deidentified to protect against observer biases. The observer mean for each parameter measured was accepted as ground truth.

By analyzing the inter- and intraobserver variability present in the manual observation process, acceptable criteria were established so as to compare the IQbot’s output to that of a trained replacement observer. Ultimately, the IQbot’s performance proved to offer significant upgrades over manual or semi-automated analysis techniques, and validation was considered successful. Further considerations ensured that the IQbot rejected and flagged concerning image-specific outputs and performed within operation time constraints.

 

The Mouse Fundus Lesion Analysis IQbot segments objects (lesions) on images of a mouse retina acquired using a fundus camera. A lesion is defined as any object that deviates from the general surrounding retinal color profile. Such deviations are generally brighter and vary in hue. These lesions can result from a number of pathologies, such as diabetic retinopathy, dry AMD, and diseases that generate focal regions of depigmentation. In order to remove any contribution of the optic nerve and vessels to the lesion count and area, the IQbot also segments these features prior to isolating, pseudocoloring, and analyzing lesions. Analysis outputs include individual and total lesion areas, lesion locations (x,y-coordinates), the optic nerve centroid, and pseudocolored overlays that provide segmentation visualization of the optic nerve, vessels, and lesions.

DEMO IMAGE
IQBOT APPLIED IMAGE

 

SAMPLE DATA
Image Total Lesion Count
Mean Lesion Area
(px)
Standard Deviation of Lesion Area
(px)
Minimum Lesion Area
(px)
Maximum Lesion Area
(px)
Total Lesion Area
(px)
Nerve Centroid X
(px)
Nerve Centroid Y
(px)
Lesion Centroid X
(px)
Lesion Centroid Y
(px)
Lesion Area
(px)
Sample 221 50.6516 304.3549 1 4502 11194 394.5433 345.5000 221 Rows of Data 221 Rows of Data 221 Rows of Data