Untitled Document

NxGen™

The NxGen™ imaging devices developed and manufactured by ImageBASE, LLC, utilizes two-dimensional image sensors capable of capturing a complete document in a single shot. This approach provides numerous advantages over conventional document scanners which are typically based on one-dimensional image sensors (fifteen year old technology).

Dramatically improved signal to noise ratios

Intermediate document modalities (film, fiche etc.) typically cover a wide gamut of quality, and conditions. Quality factors are exposure level, exposure uniformity and contrast. Condition factors are scratching, smudging and flaking emulsion. Capturing all the information stored on a photographic medium from dark or heavily exposed areas, requires a sensor system with a wide dynamic range and a low noise level.

Many commercial scanners suffer from noise or "pick-up" in the low light level environment. This condition may be due either to the engineering implementation of the imaging subsystem of the scanner or from the inherent nature of the sensors involved.

NxGen™ technology leverages the extensive amount of Research and Development effort that has been applied to the current generation of CCD and CMOS image sensors to achieve dramatically improved noise levels in the low light level environment. A key aspect of this improvement is the "on-board" electronics that have been designed into the image sensor chips. This technology typically involves preamplifier circuitry and analog to digital conversion on-chip so that all of the sensitive signal processing is done before information reaches the noisy environment of the printed circuit card.

High speed data transfer

The image sensors used in NxGen™ technology employ the "Fire Wire" (IEEE 1394) interface. This approach provides higher speed data transfer, simpler interface hardware, longer cable runs, and simpler software than the much older SCSI interface that has traditionally been the interface of choice for scanners.

True grayscale and color capability

The linearity, dynamic range and color capability of the 2-dimensional image sensors used in NxGen™ Technology provides true 8-bit (from a 10-bit A/D conversion on the sensor) grayscale and 24-bit color operation. This is contrasted to conventional commercial scanners which frequently provide no gray scale access or are often labeled as gray scale scanners but in reality provide a histogram-skewed gray.

The access to a true grayscale is extremely important when attempting to convert scans of sub-optimal quality images to the CCITT Group IV file format. Color capability is of minimal utility for filmed documents but can provide significantly enhanced image information capture in applications such as first generation digitizing of ledger books, plats etc.

Developed and maintained in-house

The NxGen™ Technology has been developed and implemented exclusively at ImageBASE, LLC. As such, the technology has been targeted specifically for the conversion of documents typical of the County Clerk environment. In addition, the maintenance aspects of the Technology are also handled in-house. This eliminates the delay involved in field service calls and spares acquisition.

No "Black Boxes"

Since the hardware and capture software have been developed at ImageBASE, LLC, there are no "Black Box" aspects of the implementation and/or operation to be dealt with. All functions, features and their operation are fully understood and maintained locally.

Rapid Customization

Given the wide range of image quality, condition and content of typical filmed document material, it is frequently desirable (perhaps necessary) to modify or tune the operation of the capture device to optimize the digitization of a particular document or family of documents. Since the hardware and software design and implementation are supported in-house, customization and new feature generation can be achieved rapidly.

Image Processing Software

Extracting information from a digital image is a complex task. ImageBASE, LLC follows the developments in image processing and document extraction in order to provide the customer with currently the best possible rendering of the information that has been on the photographic medium.

Image Enhancement

Image information stored on a photographic medium is frequently buried in blur or poor focus is often contained in density regions that are not conducive to readability or ICR/OCR operations. Applications of intelligent, region dependent contrast enhancement and level shifting, and sharpening (unsharp mask, edge enhancement and specific modification kernels) cast the digital image into a more readable or processable form.

Adaptive Thresholding Technology™

The conversion of a gray scale scan to a bitonal CCITT Group IV format involves the discarding of a large volume of information that had been stored on the photographic medium. The objective of a good conversion is to discard the information that is irrelevant to the document and to convert in a very readable fashion the relevant information. Typically, this process is performed by choosing a threshold in the gray scale and mapping all values above the threshold to white and all values below the threshold to black. The success of this process is predicated on the assumption that the original paper document was uniformly illuminated when it was filmed and that there are not extremes in the density of the information printed on the original document.

These assumptions are frequently not met and, as a result, the consequent CCITT Group IV document does not exhibit the readability that is desired by the customer and users. A proprietary element of theNxGen™ implementation is Adaptive Thresholding Technology™. This technology makes use of the latest in Neural Networks and Fuzzy Logic to compensate poor image quality resulting from poor or non uniform illumination during the filming and for variations in the density and color of information in the original document. Adaptive Thresholding Technology™ therefore produces documents of improved readability and leads to better success in subsequent ICR/OCR operations.

Targeted OCR and ICR

Some document conversion activities associated with County records require the conversion of the digital image information to a machine-readable alpha format. A typical case is the information contained in the Real estate transactions index where layouts are well defined and fonts, point sizes and character sets are limited. In this case general-purpose character recognition is saddled with unnecessary overhead and typically results in unacceptably high error rates. However, if a character recognition process is tuned to the layout, font, font point size and font character set, markedly higher recognition rates can be achieved. If, in addition, knowledge about the information content (dictionaries, name dictionaries, date fields, redundancies etc.) is known even more improvement in recognition rates can be realized.

NxGen™ technology accesses a range of recognition technologies through source code, APIs and SDKs to target recognition algorithms to the specific types and ranges of customer source material. Following conversion, databases and fuzzy logic enhanced search and matching routines are employed to build the indexes needed for managing large volume of associated document images.


	NxGen™ The NxGen™ imaging devices developed and manufactured by ImageBASE, LLC, utilizes two-dimensional image sensors capable of capturing a complete document in a single shot. This approach provides numerous advantages over conventional document scanners which are typically based on one-dimensional image sensors (fifteen year old technology). Dramatically improved signal to noise ratios Intermediate document modalities (film, fiche etc.) typically cover a wide gamut of quality, and conditions. Quality factors are exposure level, exposure uniformity and contrast. Condition factors are scratching, smudging and flaking emulsion. Capturing all the information stored on a photographic medium from dark or heavily exposed areas, requires a sensor system with a wide dynamic range and a low noise level. Many commercial scanners suffer from noise or "pick-up" in the low light level environment. This condition may be due either to the engineering implementation of the imaging subsystem of the scanner or from the inherent nature of the sensors involved. NxGen™ technology leverages the extensive amount of Research and Development effort that has been applied to the current generation of CCD and CMOS image sensors to achieve dramatically improved noise levels in the low light level environment. A key aspect of this improvement is the "on-board" electronics that have been designed into the image sensor chips. This technology typically involves preamplifier circuitry and analog to digital conversion on-chip so that all of the sensitive signal processing is done before information reaches the noisy environment of the printed circuit card. High speed data transfer The image sensors used in NxGen™ technology employ the "Fire Wire" (IEEE 1394) interface. This approach provides higher speed data transfer, simpler interface hardware, longer cable runs, and simpler software than the much older SCSI interface that has traditionally been the interface of choice for scanners. True grayscale and color capability The linearity, dynamic range and color capability of the 2-dimensional image sensors used in NxGen™ Technology provides true 8-bit (from a 10-bit A/D conversion on the sensor) grayscale and 24-bit color operation. This is contrasted to conventional commercial scanners which frequently provide no gray scale access or are often labeled as gray scale scanners but in reality provide a histogram-skewed gray. The access to a true grayscale is extremely important when attempting to convert scans of sub-optimal quality images to the CCITT Group IV file format. Color capability is of minimal utility for filmed documents but can provide significantly enhanced image information capture in applications such as first generation digitizing of ledger books, plats etc. Developed and maintained in-house The NxGen™ Technology has been developed and implemented exclusively at ImageBASE, LLC. As such, the technology has been targeted specifically for the conversion of documents typical of the County Clerk environment. In addition, the maintenance aspects of the Technology are also handled in-house. This eliminates the delay involved in field service calls and spares acquisition. No "Black Boxes" Since the hardware and capture software have been developed at ImageBASE, LLC, there are no "Black Box" aspects of the implementation and/or operation to be dealt with. All functions, features and their operation are fully understood and maintained locally. Rapid Customization Given the wide range of image quality, condition and content of typical filmed document material, it is frequently desirable (perhaps necessary) to modify or tune the operation of the capture device to optimize the digitization of a particular document or family of documents. Since the hardware and software design and implementation are supported in-house, customization and new feature generation can be achieved rapidly. Image Processing Software Extracting information from a digital image is a complex task. ImageBASE, LLC follows the developments in image processing and document extraction in order to provide the customer with currently the best possible rendering of the information that has been on the photographic medium. Image Enhancement Image information stored on a photographic medium is frequently buried in blur or poor focus is often contained in density regions that are not conducive to readability or ICR/OCR operations. Applications of intelligent, region dependent contrast enhancement and level shifting, and sharpening (unsharp mask, edge enhancement and specific modification kernels) cast the digital image into a more readable or processable form. *Adaptive Thresholding Technology™ The conversion of a gray scale scan to a bitonal CCITT Group IV format involves the discarding of a large volume of information that had been stored on the photographic medium. The objective of a good conversion is to discard the information that is irrelevant to the document and to convert in a very readable fashion the relevant information. Typically, this process is performed by choosing a threshold in the gray scale and mapping all values above the threshold to white and all values below the threshold to black. The success of this process is predicated on the assumption that the original paper document was uniformly illuminated when it was filmed and that there are not extremes in the density of the information printed on the original document. These assumptions are frequently not met and, as a result, the consequent CCITT Group IV document does not exhibit the readability that is desired by the customer and users. A proprietary element of theNxGen™ implementation is Adaptive Thresholding Technology™. This technology makes use of the latest in Neural Networks and Fuzzy Logic to compensate poor image quality resulting from poor or non uniform illumination during the filming and for variations in the density and color of information in the original document. Adaptive Thresholding Technology™ therefore produces documents of improved readability and leads to better success in subsequent ICR/OCR operations. Targeted OCR and ICR* Some document conversion activities associated with County records require the conversion of the digital image information to a machine-readable alpha format. A typical case is the information contained in the Real estate transactions index where layouts are well defined and fonts, point sizes and character sets are limited. In this case general-purpose character recognition is saddled with unnecessary overhead and typically results in unacceptably high error rates. However, if a character recognition process is tuned to the layout, font, font point size and font character set, markedly higher recognition rates can be achieved. If, in addition, knowledge about the information content (dictionaries, name dictionaries, date fields, redundancies etc.) is known even more improvement in recognition rates can be realized. NxGen™ technology accesses a range of recognition technologies through source code, APIs and SDKs to target recognition algorithms to the specific types and ranges of customer source material. Following conversion, databases and fuzzy logic enhanced search and matching routines are employed to build the indexes needed for managing large volume of associated document images.