The patent's assignee for patent number 8731260 is
News editors obtained the following quote from the background information supplied by the inventors: "The present invention relates generally to the field of virtual microscopy and pertains more specifically to data management for very large digital imaging files captured by a high resolution linear-array-based microscope slide scanner.
"Conventional scanners typically digitize a region of a physical specimen at a desired resolution. As the desired resolution increases, the scanning process becomes more technically challenging. Similarly, the scanning process becomes more challenging as the region of interest increases or as the available scanning time decreases. Furthermore, the efficiency with which the digitized data can be viewed on a monitor is often critical to the overall utility of conventional scanning applications.
"Recent technical advances in conventional sensors, computers, storage capacity, and image management have made it possible to digitize an entire microscope slide at diagnostic resolution, which is particularly desirable. Diagnostic resolution is the resolution required for a trained technician or clinician to make a diagnosis directly from a computer monitor, rather than making a diagnosis by looking through the eyepieces of a conventional microscope. Diagnostic resolution varies by sample type, for example, the diagnostic resolution required for a skin biopsy specimen is typically lower (i.e., diagnosis requires a lower resolution) than the diagnostic resolution required for other types of biopsy specimens.
"Although now technically possible, digitizing an entire microscope slide at a diagnostic resolution remains a formidable challenge. Any practical solution must capture immense amounts of high quality imagery data in a relatively short amount of time. FIG. 1 is a graph diagram plotting the limiting resolution in micrometers ('.mu.m') of an optical system with realistic condenser settings versus the numerical aperture ('NA') for the optical system's microscope objective lens. The limiting resolution is defined as the smallest distance that can be resolved by the optical system. For example, in an optical system that is designed and manufactured appropriately, the limiting resolution would be the minimum spatial dimension that can be observed by the human eye.
"As shown in the graph, the limiting resolution for an objective lens with a 0.3 NA is approximately 1.5 .mu.m. Moreover, the limiting resolution for an objective lens with a 0.4 NA improves to about 1 .mu.m while the limiting resolution for an objective lens with a 0.8 NA improves to an even better 0.5 .mu.m. At this juncture, it is important to note that the limiting resolution is independent of magnification and depends solely on the numerical aperture of the objective lens.
"Conventional systems that digitize a microscope specimen without losing any details available to the human eye require the dimension of a detector element to be no larger than one half the corresponding limiting resolution distance. This 2-pixel requirement is based on the well-known Nyquist sampling theorem. It should be clear that for a 2-dimensional imaging system, the 2-pixel requirement translates into an array of 2 pixels by 2 pixels. Stated differently, if the limiting resolution is 1 .mu.m, then it is necessary to digitize the specimen at 0.5 .mu.m per pixel (or better) to capture all of the information that is available to the human eye through the objective lens.
"FIG. 2 is a graph diagram plotting the scanning resolution in pixels per inch ('ppi') versus the numerical aperture of an objective lens. As shown in the graph, an objective lens with a 0.3 NA requires a scanning resolution of at least 38,000 ppi. This resolution is required to capture all of the details provided by the 0.03 NA objective lens and viewable by the human eye. Similarly, an objective lens with a 0.4 NA requires a scanning resolution of at least 50,000 ppi while an objective lens with a 0.8 NA requires a scanning resolution of at least 100,000 ppi.
"FIG. 3 is a graph diagram plotting the scanning resolution in pixels per inch versus the resulting uncompressed file size in megabytes ('MB') for a one square millimeter ('mm') region. The graph pertains to regions captured as 24-bit pixels (3 color channels, 8-bits per channel). As illustrated, a 1 mm2 region at 38,000 ppi is approximately 8 MB (as captured by an objective lens with a 0.03 NA according to FIG. 2). Similarly, a higher scanning resolution of 50,000 ppi for the same 1 mm2 region would result in a file size of 11 MB while a scanning resolution of 100,000 ppi would result in a file size of approximately 47 MB. As can be seen, the size of the image file increases dramatically as the required scanning resolution, expressed in pixels per inch, increases in relation to the increasing numerical aperture of the objective lens. Thus, as the scanning resolution increases, the image file size increases significantly.
"Accordingly, digitizing an entire microscope slide at a diagnostic resolution results in extremely large data files. For example, a typical 15 mm.times.15 mm slide region at a scanning resolution of 50,000 ppi (i.e., 0.4 NA) would result in a file size of approximately 2.5 gigabytes ('GB'). At a scanning resolution of 100,000 ppi, the resulting file size quadruples to approximately 10 GB for the same 225 square millimeter area of a slide.
"There are two basic methods that have been developed for scanning entire microscope slides: (i) conventional image tiling, and (ii) a novel line-scanning method and system developed by
"Conventional image tiling is a well-known technique. Image tiling involves the capture of multiple small, statically sized regions of a microscope slide using a traditional fixed-area Charge-Coupled-Device ('CCD') camera, with each capture tile being stored as a separate individual image file. Subsequently, the various image tiles that comprise a specimen are digitally 'stitched' together (i.e., alignment) to create a large contiguous digital image of the entire slide.
"The number of individual image tiles required to scan a given area of a slide is proportional to the number of pixels that comprise each image tile. A typical video-format color camera has 768.times.494 pixels, which translates into 1.1 MB of imagery data per image tile. Recalling that a 1 mm2 region of a slide corresponds to 11 MB of imagery data, it follows that approximately 10 non-overlapping image tiles must be captured to digitize one square millimeter of a slide at a scanning resolution of 50,000 ppi. At 100,000 ppi the required number of tiles increases four-fold to 40 image tiles per square millimeter.
"It follows that for a typical 15 mm.times.15 mm slide region, at a scanning resolution of 50,000 ppi, a minimum of 2,250 individual image tiles must be captured. At a scanning resolution of 100,000 ppi, a minimum of 9,000 individual image tiles must be captured. Importantly, each image tile would have a file size of approximately 1.1 MB. In practice, an even larger number of tiles must be captured to provide sufficient overlap between adjacent tiles to facilitate the 'stitching' together or alignment of adjacent image tiles.
"Conventional image tiling systems generally take hours to capture and align the thousands of tiles required to digitize an entire microscope slide. Image capture times are significantly increased by the need to wait for the CCD camera to stabilize after being repositioned and before acquiring an image tile. This wait time is necessary to ensure that the captured image does not blur. Practical limitations in data processing speeds also make the alignment of large numbers of image tiles extremely slow. In practice, conventional image tiling systems are not able to align large numbers of tiles without creating 'stitch lines' and other image artifacts that create computer imaging challenges.
"An alternative to image tiling is the afore-mentioned line-scanning method. Rather than using a fixed-area camera to capture thousands of individual image tiles, the line-scanning method employs a linear-array detector in conjunction with a microscope objective lens and other optics to capture a small number of contiguous overlapping image stripes. Unlike the stop-and-go nature of conventional image tiling, the microscope slide moves continuously and at a constant velocity during acquisition of an image stripe. One of the many fundamental advantages of line-scanning over conventional image tiling is that the capture and alignment of a small number of image stripes is significantly more efficient than the capture and alignment of thousands of separately captured image tiles.
"For example, a typical 15 mm.times.15 mm slide region at 50,000 ppi would require 15 image stripes, each with a width of 2,000 pixels, to digitally capture the region. Here, each image stripe would have a file size of approximately 170 MB. At 100,000 ppi, the same region would require 30 image stripes with each stripe comprising approximately 680 MB. The capture of 15 or 30 image stripes for a 15 mm.times.15 mm area is dramatically more efficient than the capture of 2,250 or 9,000 image tiles at 50,000 ppi or 100,000 ppi respectively. Furthermore, the continuous scanning nature of line-scanning makes it possible to create seamless virtual slides of a region in minutes.
"In addition to rapid data capture, line scanning benefits from several advantages that ensure consistently superior imagery data. First, it is possible to adjust the focus of the objective lens from one scan line to the next, in contrast to image tiling systems that are inherently limited to a single focal plane for an entire image tile. Second, because the sensor in a line scanning system is one-dimensional, there are no optical aberrations along the scanning axis. In an image tiling system, the optical aberrations are circularly symmetric about the center of the image tile. Third, the linear detector has a one-hundred percent (100%) fill factor, providing full pixel resolution (8 bits per color channel), unlike color CCD cameras that lose spatial resolution because color values from non-adjacent pixels are interpolated (e.g., using a Bayer Mask).
"To handle the immense amounts of data produced by conventional image tiling systems, data management tools have been developed to manage the thousands of relatively small (.about.1 MB) image tiles typically generated by such systems. These data management utilities, however, are not suitable for managing a small number of relatively large (.about.200 MB) image stripes captured by the line-scanning image striping system.
"Therefore, introduction of the superior image striping system and method for digitizing microscope slides has created a need in the industry for a data management system that meets the unique needs imposed by the new technology."
As a supplement to the background information on this patent, VerticalNews correspondents also obtained the inventors' summary information for this patent: "The present invention provides a data management system and method for processing and handling the extremely large imagery data files (i.e., image stripes) that are rapidly produced by a linear-array-based microscope slide scanner. The system receives, processes, and stores the high volume of imagery data, which is produced by the linear-array-based microscope slide scanner at approximately 3 GB per minute.
"The data are received as a series of coarsely aligned, slightly overlapping image stripes that are corrected for image non-uniformities and chromatic aberrations and then finely aligned into a seamless and contiguous baseline image. The baseline image is then logically mapped into a plurality of regions that are individually addressed to facilitate viewing and manipulation of the baseline image. These plurality of regions are referred to in the industry as 'image tiles' but should not be confused with the various separate image files that are individually captured by a CCD camera in a conventional image tiling system.
"The data management system enables imagery data compression while scanning and capturing new image stripes. This advantageously eliminates the overhead associated with storing uncompressed image stripes. The image compression also creates intermediate level images, thereby organizing the baseline image into a variable level pyramid structure referred to as a virtual slide.
"The data management system efficiently converts image stripes into a high quality virtual slide that allows rapid panning and zooming by image viewing software. The virtual slide also allows efficient processing by an algorithm framework. Furthermore, the functions of real-time image processing, compression, and storage are combined with simultaneous and simplified multi-resolution viewing of high quality images at local and remote stations. The data management system is cost effective and scaleable, employs standard image file formats and supports the use of virtual slides in desirable applications such as telemedicine, telepathology, microscopy education, and the analysis of high value specimens such as tissue arrays."
For additional information on this patent, see: Crandall, Greg J.; Eichhorn, Ole; Olson, Allen H.; Soenksen, Dirk G.. Data Management in a Linear-Array-Based Microscope Slide Scanner. U.S. Patent Number 8731260, filed
Keywords for this news article include: Information Technology,
Our reports deliver fact-based news of research and discoveries from around the world. Copyright 2014, NewsRx LLC
Most Popular Stories
- Stop-Start Engines Save Gas, Reduce Emissions
- Shia LaBeouf Plea Deal, Alcoholism Treatment
- Ohio State Band Chief Fired After Probe
- Hispanic Leader Goes the Extra Mile
- World Tensions Don't Curb Enthusiasm for Stocks
- U.K. Economy Surpasses Pre-Crisis Peak
- Morgan Stanley Ponies Up $275 Million to Settle SEC Charges
- Ricky Martin Joins 'The Voice ... Mexico'
- Jennifer Lopez, Pitbull to Perform at Fashion Rocks
- Visa, Amazon Results Drag Down the Street