Instrument Insider
The Megapixel Myth
A fundus camera's MP number has little bearing on image quality.
By Scott Grant, M.D.
Anyone in the market for a fundus camera will quickly become aware of the supposed preeminence of megapixels. Many camera companies promote the number of megapixels in their images as integral to their camera's quality.
In reality, this is only a small aspect of image quality and may be just part of marketing strategy. Larger pixel counts do not always translate into better image resolution. Unless you intend to print giant, wall-size posters of your patients' retinas, 3 to 5 million pixels is usually adequate for good imaging, and anything over 11 million is much more than necessary. Yet some marketing claims tout cameras with 24 million pixels, which is silly for standard retinal imaging.
As a retinal specialist with years of experience in fundus film photography, I realize that megapixels are a small part in the quality of images one gets with fundus photography.
The Elements of Image
Images on a computer are composed of tiny dots called pixels (a word derived from the term PICture ELement) arranged horizontally and vertically on the screen. In stored image files, resolution is determined by how many pixels run horizontally or vertically.
Pixel count expressed in megapixels (MP) is the value obtained by multiplying the number of horizontal pixels by the number of vertical pixels, the same way you calculate square inches of an area. Thus an image with 2,048 horizontal pixels by 1,536 vertical pixels has 3,145,728 pixels or approximately 3 megapixels.
Because megapixel counts are square functions, it takes very large differences in MP to translate into significant linear pixel difference. For example, a 3 MP camera has 2,048 pixels horizontally and a 14 MP camera has 4,500. Hence, a nearly five-fold increase in megapixel power only raises horizontal resolution by a little over half.
By promoting megapixels as a marketing strategy, fundus camera makers are following the lead of their counterparts in the consumer world. Over the last decade, digital consumer cameras have struggled to outdo each other in MP power, sometimes at the expense of image quality, as professional photographers have pointed out. To increase megapixels but still maintain their small size, consumer cameras have been forced to squeeze more and more photosites on to their image chips. These smaller photosites receive less light and are subject to more noise and other aberrations due to their tightly packed arrangement. These noisy, dim photo captures then must be artificially boosted and denoised before being written to the camera's flashcard. "You end up with a photograph that is technically higher-resolution," observes an editorial in Digital Camera magazine, "but may in fact be of inferior subjective quality compared to the same shot taken with a camera offering a more modest pixel count."1
Empty Magnification
Image quality in fundus photography is primarily determined by the interaction of the fundus camera and the optical system of the eye. Lens quality, camera design issues, filter quality and the optical quality of the particular eye in question are the primary factors in the image quality. Once the image is defined by the camera/eye system, the digital film comes into play to record the picture. Too many pixels will not cause you to miss a diagnosis or misinterpret the patient's status. But, on the other hand, excessively large photo images occupy more storage space and can slow your system down. Additionally, if you judge the quality of digital fundus camera based primarily on pixel count, you are not accurately evaluating that device.
Empty magnification is a concept discussed in microscope design. It means that even if you enlarge an image, the detail visible will not be fundamentally greater. If you can clearly see a capillary in an image, without the jagged, pixilated edge of low resolution, enlarging the image by another 10 times will not permit you to view the cells that make up that capillary — all you will be getting is a bigger picture of a capillary. Early-1980s film audiences were wowed by a scene in Blade Runner in which Harrison Ford's detective repeatedly magnifies a photo until it reveals a key piece of evidence not visible in the original image, but such capabilities are still science fiction at this point.
Even the least expensive consumer digital camera employs very complex scientific principles. It is no surprise that consumers and manufacturers have been eager to grasp at the number of megapixels as an easy-to-understand way to differentiate among cameras. But as much as we would like to boil down all imaging complexity into one simple feature like pixel count, the fact is no such yard-stick exists. Lens quality, even in consumer cameras, is still the most important aspect of the camera determining image quality.
Having used fundus cameras for over 30 years and bridging the film to the digital era, I still am unable to give a simple answer in picking one camera system over another.
However, I can offer one basic rule of thumb: the quality of the camera's lens and optical system is more important than the number of pixels in the image produced.
There are some basic qualities to look for. For instance, the camera should be able to obtain clear images through cloudy media (cataract), and the lens should admit as much light as possible. The images should be sharp but also have good contrast. The digital color depth should be high to allow for natural true color.
Ultimately, the only way to tell if a camera can capture quality images is to use it yourself. Do not rely only on samples provided by the company, since they were produced under optimal conditions. Image several different kinds of patients under as many different scenarios as possible.
It has only been a short time since digital imaging has displaced film. It has provided a wonderful technology to instantly view images and electronically manipulate them. Showing a patient what is being visualized is an excellent tool for patient education. So pick a system that is user friendly and takes great pictures, knowing that pixels are just one of the many variables to consider. OM
Reference
1. David MacNeill. "10 megapixels: Too much of a good thing?" Digital Camera. April 2006:10;39.
Scott Grant, M.D., is a retinal surgeon who practices at the Retinal Consultants of Orange County, which has offices in Fullerton and Los Alamitos, Calif. He has an under-graduate degree in physics and a masters degree in Applied Physics. |