Effect of Image Compression and Resolution on Retinal Vascular Caliber

Abstract

Purpose.: Changes in retinal vascular caliber measured from digital color fundus photographs have been independently associated with systemic outcomes in epidemiologic studies, but the effect of image resolution and compression on vascular measurements has not been previously evaluated.

Methods.: To explore image compression, 40 natively digital fundus images were selected with good photo quality, high spatial resolution, and no previous image compression. Using Adobe Photoshop, these images were compressed at progressively higher levels up to 147:1, and then retinal vascular caliber was measured at each level using semiautomated software. To examine resolution, 40 fundus photographs acquired on high-resolution film were scanned with settings corresponding to 10, 7, 5, 3, and 1 megapixel fundus cameras. After adjusting for scale factor, vascular caliber was measured at each level of resolution. Data were analyzed by comparing the calculated central retinal arteriole equivalent (CRAE) and the central retinal venular equivalent (CRVE) of the original and altered images, using repeated measures ANOVA.

Results.: CRAE became significantly wider with increasing levels of compression at the 25:1 threshold (∼1 μm wider, P < 0.001) and was ∼5 μm wider with 147:1 compression. CRVE also increased, but less than CRAE. Using 7 (megapixel)-MP resolution as the standard, CRVE was significantly narrower at the 5-MP simulation (∼2 μm, P < 0.001) and was ∼12 μm narrower at the 1-MP simulation. CRAE also decreased, but less than CRVE.

Conclusions.: Increasing digital image file compression and decreasing fundus image spatial resolution led to skewed measurements of the retinal vascular caliber.

Introduction

Epidemiologic studies have reported that retinal vascular caliber changes are independent risk factors for different systemic health outcomes. Collectively, this methodology has been applied to the images of more than 40,000 individuals. ¹ Commonly, retinal vascular caliber is measured from digital color fundus images—natively digital or scanned from film—using specialized software that detects vessel edges in order to calculate the width of arterioles and venules. For each eye, these individual vessel widths are then combined using a generalized model (Knudtson revision of the Parr-Hubbard formulas) to obtain the central retinal arteriolar equivalent (CRAE) and the central retinal venular equivalent (CRVE). ²  

The uncompressed files of digital images may be large—on the order of 14 MB for a 5- MP camera. To facilitate storage and transportation of large images, they are often compressed to reduce file size. However, any compression beyond ∼4:1 requires the use of “lossy” (as opposed to lossless) compression, typically in the joint picture expert group (JPEG) format. Such compression increasingly tends to diminish image detail as the file size decreases. Thus, it is critically important to know how much the image is compressed as the file size decreases, to ensure that sufficient image integrity is preserved for accurate measurement of the retinal vessels. 

Closely related is the issue of spatial resolution—in digital images the x × y pixel dimensions of the image, expressed in megapixels. The color slide transparency emulsion used historically, Ektachrome or equivalent (Eastman Kodak Company, Rochester, NY), has been estimated to correspond to the spatial resolution of a 10-MP digital image, approximately 3000 × 3600 pixels. ³ Regardless of digital image compression, various digital fundus camera systems capture images at different pixel dimensions, which also impacts file size. Two primary parameters of digital fundus cameras that affect the spatial resolution of the digital image are the degree of magnification of the image provided by the camera optics, and the dimensions of the pixel array of the digital sensor itself. However, the size of the sensor pixel array available to capture the image is fixed regardless of the optical field setting. Thus, from any given camera, wide-angle images at 45° to 50° have lower spatial resolution than do approximately 30° images. Because the variety of digital fundus cameras and digital sensors now in use is greater than the variety of historical film cameras, it is prudent to be cognizant of the possible impact of spatial resolution of the digital fundus image on retinal vascular caliber measurements. 

To our knowledge, progressive image compression has not been studied in the context of retinal vascular caliber measurement. Nor are we aware of any research on the possible impact of differential image resolution on accuracy of retinal vascular measurement. 

Methods

Images from 40 participant eyes were randomly selected from multiple studies within the University of Wisconsin-Madison Fundus Photograph Reading Center database. These studies had been conducted in accordance with the tenets of the Declaration of Helsinki and under supervision of the respective institutional review boards. All participants had given written informed consent to have their fundus photographs taken for research purposes. Using Irfanview image handling software (Version 4.27; Irfan Skiljan, Vienna, Austria), the image information was viewed to ensure that the image had no previous compression. Furthermore, the image evaluator had to deem the image focus and illumination to be adequate. All photos had been taken with digital cameras manufactured by either Topcon (Topcon Medical Systems Inc., Oakland, NJ) or Zeiss (Carl Zeiss Meditec Inc., Dublin, CA). The average file size of the original uncompressed tagged image file format (TIFF) image was approximately 14 MB. 

For vascular measurement, the “scaling” factor or the “pixel spacing” parameter (the Digital Imaging and Communications in Medicine [DICOM] standard; National Electrical Manufacturers Association, Rosslyn, VA, 2009), of each image is determined prior to evaluation. The number of pixels across a known distance (disc to macula) is evaluated, and the pixel spacing parameter is calculated by dividing the number of pixels by the known distance. In this study, while the original image is down-sampled for analysis, the pixel spacing parameter is unchanged by image compression but altered as spatial resolution changes. We used the landmarks of the center of the disc and the center of the macula, which is presumed to be 4500 μm, to determine the pixel spacing value. 

Predominantly, retinal imaging in clinical settings uses JPEG compression. ⁴ Accordingly, for this exercise the uncompressed TIFF images were imported into Adobe Photoshop (CS3 V. 10.0; Adobe Systems Inc., San Jose, CA) and then converted to JPEG images at compression ratios ranging from 4:1 to 147:1. The compressed JPEG files were then converted back to TIFF format as required to use the vessel measurement software, IVAN (Version 1.3; Nicola Ferrier, Madison, WI). For each camera and angle combination, an average pixel spacing factor was determined across all the original images. Image compression does not change the pixel spacing factor, thus the same factor is applicable at all compression ratios. To eliminate interobserver variability, one person evaluated all the images in this study, with the original image being evaluated first. Before recording the data of the original image, a screenshot was saved as a reference for grid placement and vessel identification, ensuring that the corresponding sections of the vessels would be measured on the subsequent compressed images. After summarization into the CRAE and CRVE, results were collected for the statistical analysis, which is described later. A flow chart of the methods used for the compression portion of the study is shown in Figure 1. 

Figure 1. 

 

Flow chart of the methods for the compression study.

The effect of progressively lower spatial resolution was assessed by analyzing measurements on images that correspond with 10-, 7-, 5-, 3-, and 1-MP cameras. To simulate different camera resolutions using the same images, we used film images from 40 eyes of participants who were randomly selected from the multiple studies within the University of Wisconsin-Madison Fundus Photograph Reading Center database. The camera types from the images selected were the Zeiss 30, Topcon 50, and Canon 60 degree film cameras (Canon U.S.A., Inc., Lake Success, NY). 

Slide images were digitized using a Nikon Co-olscan V ED 1.02 scanner (Nikon Inc., Melville, NY) at the pixel dimension calculated to represent the different camera spatial resolutions listed previously. Changing the resolution of an image alters the pixel spacing parameter; therefore, we had to establish the corresponding pixel spacing parameter for each subsequent resolution. As each series was created by altering an original slide, the millimeter distance from the center of the disc to the center of the macula remained unchanged, as the only factor varying through the series was the pixel spacing parameter in each image representing this distance. With this knowledge, a calibration ratio was created that could be used to scale each image relative to others in the set. Therefore, by measuring the distance between the center of the disc to the center of the macula and the distance across the entire image in pixels, the ratio can be calculated for each image. The average ratio from the five images was used to calculate the calibration ratio for that image set. The calibration ratio was then implemented by multiplying it against the measured distance across each image. This yielded a disc to macula measurement that was consistent across all spatial-resolution images in millimeters regardless of pixel spacing, ensuring accurate measurement and grid fit. The 7-MP images were evaluated first, on a monitor with a display resolution of 1600 × 1200 pixels, with the pixel spacing changed accordingly within the configuration file before evaluating the other images. Similar to the compression portion of this exercise, we saved a screenshot of the original image to use as a reference for subsequent images. A flow chart of the methods for the spatial resolution portion of the study is shown in Figure 2. 

Figure 2. 

 

Flow chart of the methods for the spatial resolution study.

In both studies, CRAE and CRVE results were compared by calculating the signed difference between the original images and each image variant in the study. In analysis of variance for repeated measures design, each image serves as its own control, and accordingly, variability owing to differences in average responsiveness of the images is eliminated from the extraneous error variances. Data management and analyses were performed in SAS (SAS/STAT software, Version 9.2; SAS Institute Inc., Cary, NC). 

Sample size calculations were performed using G*Power (Version 3.1.3; Franz Faul, Kiel University, Kiel, Germany). The sample size required to demonstrate a difference within two compression levels was 27 images (α = 0.05, power = 0.80, and effect size = 0.1), while the sample size for resolution was 38 images (α = 0.05, power = 0.80, and effect size = 0.1). 

Retinal vascular caliber is routinely measured at the reading center for other observational cohorts, and we perform reproducibility exercises as part of the quality control process. Historically we have observed that variability within evaluator, the relevant consideration in this experiment, is less than that between evaluators. 

Results

Both arterioles and venules appear progressively broadened with greater image compression (Tables 1 and 2; Fig. 3) with arterioles affected more at each compression ratio. The differences between the vessel diameter measured in the original image and in the compressed image become statistically significant for CRAE at the 25:1 compression ratio (P = 0.0003) and for CRVE at the 42:1 compression ratio (P = 0.03). 

Figure 3. 

 

Vessel broadening as the image becomes more aggressively compressed.

Table 1. 

 

CRAE Mean Difference Comparing Original Uncompressed Image (1:1) with Decreasing Compression Ratios (x:1) (n = 40)

Compression Ratio	File Size (MB)	Estimate	Standard Error	Pr > |0.05|*	Lower	Upper
4:1	3.82	−0.01	0.22	0.97	−0.45	0.43
11:1	1.20	0.34	0.24	0.15	−0.13	0.82
25:1	0.55	0.85	0.21	0.00	0.42	1.28
42:1	0.33	1.39	0.32	<0.0001	0.76	2.03
60:1	0.23	1.95	0.27	<0.0001	1.41	2.48
88:1	0.16	3.03	0.31	<0.0001	2.40	3.66
115:1	0.12	2.65	0.37	<0.0001	1.89	3.40
147:1	0.09	4.84	0.49	<0.0001	3.85	5.83

  * P > |0.05|

Table 2

 

CRVE Mean Difference Comparing Original Uncompressed Image (1:1) with Decreasing Compression Ratios (x:1) (n = 40)

Compression Ratio	File Size (MB)	Estimate	Standard Error	Pr > |0.05|	Lower	Upper
4:1	3.82	0.28	0.24	0.25	−0.21	0.77
11:1	1.20	−0.03	0.17	0.87	−0.38	0.32
25:1	0.55	0.39	0.24	0.12	−0.10	0.88
42:1	0.33	0.60	0.27	0.03	0.05	1.15
60:1	0.23	0.91	0.32	0.01	0.27	1.55
88:1	0.16	1.38	0.39	0.00	0.59	2.16
115:1	0.12	1.53	0.39	0.00	0.74	2.32
147:1	0.09	2.84	0.37	<0.0001	2.09	3.60

Decreasing spatial resolution produced the artifact of narrowing measurements compared with each of the scan resolutions used (Tables 3 and 4; Fig. 4), with the venules distorted at a higher rate of change in width measurements than the arterioles. Using the 7-MP scanned image as the standard, there is a significant difference for both CRAE and CRVE at every camera resolution that was simulated (P < 0.001). 

Figure 4. 

 

Apparent vessel narrowing as spatial resolution is decreased.

Table 3. 

 

CRAE Mean Difference Comparing Reference Image (7 MP) and the Corresponding Spatial Resolution (n = 40)

Spatial Resolution	Estimate	Standard Error	Pr > |0.05|	Lower	Upper
1 MP (1024 × 768)	−7.25	0.94	<0.0001	−9.14	−5.35
3 MP (1872 × 1603)	−2.77	0.29	<0.0001	−3.35	−2.18
5 MP (2392 × 2048)	−1.57	0.26	<0.0001	−2.10	−1.04
10 MP (3383 × 2896)	0.74	0.26	0.01	0.21	1.27

Table 4. 

 

CRVE Mean Difference Comparing Reference Image (7 MP) and the Corresponding Spatial Resolution (n = 40)

Spatial Resolution	Estimate	Standard Error	Pr > |0.05|	Lower	Upper
1 MP (1024 × 768)	−15.26	2.82	<0.0001	−20.96	−9.56
3 MP (1872 × 1603)	−4.89	0.46	<0.0001	−5.82	−3.95
5 MP (2392 × 2048)	−1.93	0.31	<0.0001	−2.56	−1.30
10 MP (3383 × 2896)	1.07	0.26	0.00	0.54	1.61

Discussion

Our study showed that small yet statistically significant change in retinal vascular measurements can be artifactually caused by higher levels of lossy image compression and by lower degrees of spatial resolution. These differences became larger with more aggressive compression (approximately 5 μm at 147:1 compression) and with very low resolution (approximately 12 μm with a 1-MP compared with a 7-MP image). Presumably, aggressive image compression has a blur effect on the edges of the vessels, which are known to cause them to appear artificially broader. This may be explained in a luminance analysis as having wider slopes as the image is further compressed. Whereas lower resolution images have less initial data displayed, owing to fewer pixels in the image, one can expect sharper slopes in a luminance analysis and therefore narrower vessels. 

Current procedures of acquiring and transmitting images between clinics, to a centralized reading center, or a coordinating center, frequently employ compression of images in order to save space and allow them to be easily transmitted. This process is also widely used in telemedicine. It has been shown in previous studies that compression ratios up to 30:1 do not significantly impair the ability of an ophthalmologist to evaluate disease status from the image. ⁵ At the University of Wisconsin-Madison Fundus Photograph Reading Center, we require all images to be submitted as TIFF, PNG, or BMP. These images have no previous compression, which allows for the best assessment of an image. It is possible to use varying techniques of digital forensics to detect if an image has been previously compressed. ^6,7  

Spatial image resolution is determined by the pixel height and width of the image, usually described in megapixels, which is constant for each camera type. There was a statistical difference between the measurements evaluated when comparing the five spatial resolutions with each other. As the resolution decreased, we noted that the apparent vessel diameter also decreased in size. Figure 5 demonstrates the process by which IVAN draws splats on the retinal vessels. The splats when grouped together help define vessel wall boundaries, and with reduced resolution, this definition is less pronounced, leading to narrower vessel measurements. This may be partially explained by the greater pixel spacing at the low resolutions. Pixel spacing with the 1-MP cameras with 50 and 60 degree of field were as high as 15 to 17 microns apart, whereas the reference image (7 MP) for the same subject had pixel spacing ranging from 4.5 to 6.5 microns. Film-based studies used 30° to 35° fixed camera angles. However, more recent cameras have the ability to range from 20° to 60° with an angle selector. Because the spatial resolution is set for a particular digital system, changes in camera angle will affect the number of pixels in the region in which the vessels are measured, which may also introduce variability. 

Figure 5. 

 

Analysis using IVAN software on a film image scanned at 10 MP (A) and the same film image scanned at 1 MP (B). The right-hand panel for each image shows the “splats” or uniform regions that the software was able to detect in each image. Compared with granularity in A, the splats are rarified in B. This difference has two consequences: the length of each vessel segment detected by the software is shortened and the width ascribed to these vessels is somewhat narrowed.

In epidemiologic studies, an increase or decrease of even a few micrometers could move a subject from the appropriate quartile into a different quartile. ^8–10 This in turn could lead to data being wrongly categorized, hence affecting analysis of studies. Therefore, the level of image compression for research studies should be chosen judiciously so as to avoid an artificial broadening of the vessels that might skew the data at a statistically significant level. 

A limitation is that we did not prospectively collect data using different digital cameras or sensors to obtain images from the same subjects at the varying spatial resolutions. Therefore the simulations employed may not translate precisely to the way digital cameras capture the retinal images. However, when we evaluated the images, no differences could be determined between the digital images and the film images scanned at the same pixel dimensions. The procedure of scanning in the same image five times eliminated the possibility of technical variability by the photographer. Previous studies show that there is some variability in vascular caliber measurement between images of the same eye even when capturing the images in a short time interval. ¹¹ Technical variables include, but are not limited to, varying focus of the subject, an image being photographed at different parts of the pulse cycle, ¹² and artifacts. By taking this approach we were able to ensure that the change in spatial resolution would be the only variable that could change between the images. 

Conclusions

Varying the spatial resolution or file compression of a digital fundus image can have a significant effect on retinal vascular caliber measurements. There are currently no accepted standards for levels of resolution and compression that would yield accurate CRAE and CRVE measurements. However, our results suggest that a compression ratio of 25:1 or greater is unacceptable for a 5-MP camera. Minimum spatial resolution is best expressed as minimum pixel spacing because of the ability to also change the degree of field. We believe this number will fall between 9 and 15 μm/pixel, which is representative of the range between our 3-MP and 1-MP cameras. 

Acknowledgments

The authors thank Yijun Huang, PhD, for sharing his expertise on digital image file types and advising us on this manuscript. 

Footnotes

 Supported by Research for Prevention of Blindness.

Footnotes

 Disclosure: T.W. Pauli, None; S. Gangaputra, None; L.D. Hubbard, None; D.W. Thayer, None; C.S. Chandler, None; Q. Peng, None; A. Narkar, None; N.J. Ferrier, None; R.P. Danis, None

Footnotes

 Reprint requests: Ronald P. Danis, Fundus Photograph Reading Center, Department of Ophthalmology and Visual Sciences, University of Wisconsin-Madison, 8010 Excelsior Drive, Suite 100, Madison, WI 53717; .