Stellaris Technologies Logo
 
customer feedback
 
FAQ | Awards | Customer Speak | Demo
STClaims | STDentalClaims | STTranscripts
STModules | STModuleTracker | STEmailClassifier
STSmartTables | DIAStatPro | STTax
 
 
STSmartFilters
Smart Image Processing Filters
 
 
     
 
STSmartFilters comprises of a suite of Image Processing Filters that can also be configured as Client modules to work within a Captiva Capture workflow.
 
     
 
 
 

STSmartFilters comprises of

  • STPixProcessor - a Filter that helps convert color and greyscale images to black & white
  • STPageReg - a module that helps with aligning images that belong to a structured form type
  • STSmartOCR - a set of cleaning filters that assist with enhancing the image making it ready for OCR
  • STPixID - a filter that helps with extracting data from identification documents such as Drivers License, Passport, and others

STPixProcessor
As Mobile Capture makes inroads into business processes, more and more documents are now captured on mobile devices by customers and remotely located employees. Traditional scanning operations were centralized and captured black and white documents using high-speed scanners. Document Capture systems evolved to get better and better at handling these documents. However, in the mobile world, the kind of documents expected to be processed have grown extremely diverse. While mobile capture simplifies the operational process, it also brings in its own set of unique challenges. Perspective errors, skews, uneven illumination, lack of clarity are just a few of them.

While out of the box Image Processing filters use static thresholding applied uniformly across all color images to convert them to black & white, we have seen that images captured using mobile devices vary in contrast, hue, intensity and color across the image itself.
STPixProcessor uses dynamic thresholding to convert images to black & white, a self correcting algorithm that adjusts the theshold for conversion to black and white on different portions of the same image. The resultant images are of better quality and more OCR friendly.
The transformation is so pronounced that there is a huge leap in accuracy. This could translate to huge savings in the Document Capture business processes, but more importantly it allows customers and remote employees to capture documents on the mobile rather quickly without requiring to know about illumination or other factors that affect document capture.


 
     
 
STPixProcessor
 
     
 

STPageReg
One of the biggest challenges in document capture is the minor variations in the size of documents that causes fields to move around rather than always fall in exactly the same place. Paper forms may vary in size to some extent depending on the printer settings that were used to print the form. When scanning paper forms, the scanners themselves introduce minor shifts and skews in the images. While this may not seem like an issue while looking at the image since the image is still just as readable for the human eye, OCR software that depends on precise zones could be affected quite significantly. This in turn causes a drop in OCR accuracy and increased operator costs for manual correction of data.

Stellaris Technologies provides STPageReg, a Captiva client module for precise registration of images. It aligns all images to a pre-selected gold image so that all images and fields fall exactly on top of each other. If held against light, one could see the forms line up precisely. Once the form images are processed through STPageReg, they can be sent to Extraction where the field data is OCRed. This module can be used on drop out and non dropout forms and has been seen to provide excellent improvements in OCR accuracy.

Including STPageReg in your Captiva workflow allows the software to give you the best possible OCR results and get the most from your Capture Workflow.

 
     
 
STPageReg
 
     
 

STSmartOCR
Capturing data from a form involves a large number of fields potentially across multiple pages, cross field dependencies and critical header information. More specifically, capturing data from forms where the labels and lines on the form continue to be part of the captured image presents a problem. The OCR accuracy on them is usually poor and a lot of time and effort is spent on keying the data. OCR engines work well with uniformly sized text. With machine printed data interspersed between pre-printed labels of a different size, OCR engines sometimes read the content erroneously.

STSmartOCR is a suite of image filters that is capable of enhancing images - structured as well as unstructured - making it more OCR friendly and hence improving OCR accuracy. STSmartOCR provides a design interface called STSmartOCRDesigner that allows the developer to select from a wide range of filters - Line Removal, Dash Removal, Comb Removal, Associated Label Removal, Edge Removal, Invert Colors, Erase Between Lines, Block Removal and others.
The designer is a very intuitive application that allows users to configure these filters and test them on sample images before incorporating them into a capture workflow. Some filters have configurable parameters that can be adjusted to obtain the best results.

STSmartOCR is known to provide breathtaking results on images where data characters lie in close proximity or on top of the lines or when the labels are too close to the data to be captured. It can also invert colors intelligently so that portions of the image that have white labels on a black background can be reversed to obtain good OCR reads and hence high accuracy results.

 
     
 
STSmartOCR
 
     
 
STSmartOCR
 
     
 

STPixID/STEnhanceOCR
With the advent of mobile technology there has been a growing demand for identifying and extracting data from Identification documents such as drivers licenses, passports, national ID cards. The images that are captured range from being color to greyscale to black and white and vary significantly in quality. The font sizes as well as the colors of the fonts and the backgrounds are highly diverse posing a lot of problems for high accuracy OCR capture. Although data to be captured is restricted to a handful of fields, the source documents being unstructured and data sitting on watermarks and patterns in the background poses a challenge. Further the images can be captured in any orientation and can be skewed in any direction.

Using a combination of filters with STPixID and STEnhanceOCR, Stellaris Technologies has been able to carve out a solution that can help in cleaning up the images as well as background, correcting the skews and extracting relevant blocks of data. The solution has been tested against many differnet types of images and has resulted in high OCR accuracy.

 
     
     
 
STSmartOCR
 
     
 
     
 
     
   
Company
Products
Expertise
Success Stories

Our Team
Reach Us
Stellaris Technologies logo
 
Facebook Tweeter linkd Youtube
 
Copyright © Stellaris Technologies Pvt. Ltd.
 
Stellaris Technologies Logo
Reach Us Our Team Success Stories Expertise Products Home
feedback
 
 
STSmartFilters
Smart Image Processing Filters
 
 
     
 
STSmartFilters comprises of a suite of Image Processing Filters that can also be configured as Client modules to work within a Captiva Capture workflow.

 
 

STSmartFilters comprises of

  • STPixProcessor - a Filter that helps convert color and greyscale images to black & white

  • STPageReg - a module that helps with aligning images that belong to a structured form type

  • STSmartOCR - a set of cleaning filters that assist with enhancing the image making it ready for OCR

  • STPixID - a filter that helps with extracting data from identification documents such as Drivers License, Passport, and others
 

 

 
 
 
STPixProcessor
As Mobile Capture makes inroads into business processes, more and more documents are now captured on mobile devices by customers and remotely located employees. Traditional scanning operations were centralized and captured black and white documents using high-speed scanners. Document Capture systems evolved to get better and better at handling these documents. However, in the mobile world, the kind of documents expected to be processed have grown extremely diverse. While mobile capture simplifies the operational process, it also brings in its own set of unique challenges. Perspective errors, skews, uneven illumination, lack of clarity are just a few of them.


 
 

While out of the box Image Processing filters use static thresholding applied uniformly across all color images to convert them to black & white, we have seen that images captured using mobile devices vary in contrast, hue, intensity and color across the image itself. STPixProcessor uses dynamic thresholding to convert images to black & white, a self correcting algorithm that adjusts the theshold for conversion to black and white on different portions of the same image. The resultant images are of better quality and more OCR friendly.


The transformation is so pronounced that there is a huge leap in accuracy. This could translate to huge savings in the Document Capture business processes, but more importantly it allows customers and remote employees to capture documents on the mobile rather quickly without requiring to know about illumination or other factors that affect document capture.

 

 

 
STPixProcessor
 
     
 
STPageReg
One of the biggest challenges in document capture is the minor variations in the size of documents that causes fields to move around rather than always fall in exactly the same place. Paper forms may vary in size to some extent depending on the printer settings that were used to print the form. When scanning paper forms, the scanners themselves introduce minor shifts and skews in the images. While this may not seem like an issue while looking at the image since the image is still just as readable for the human eye, OCR software that depends on precise zones could be affected quite significantly. This in turn causes a drop in OCR accuracy and increased operator costs for manual correction of data.
 
 

Stellaris Technologies provides STPageReg, a Captiva client module for precise registration of images. It aligns all images to a pre-selected gold image so that all images and fields fall exactly on top of each other.

If held against light, one could see the forms line up precisely. Once the form images are processed through STPageReg, they can be sent to Extraction where the field data is OCRed. This module can be used on drop out and non dropout forms and has been seen to provide excellent improvements in OCR accuracy.

Including STPageReg in your Captiva workflow allows the software to give you the best possible OCR results and get the most from your Capture Workflow.

 

 

 
STPageReg
 
     
  STSmartOCR
Capturing data from a form involves a large number of fields potentially across multiple pages, cross field dependencies and critical header information. More specifically, capturing data from forms where the labels and lines on the form continue to be part of the captured image presents a problem. The OCR accuracy on them is usually poor and a lot of time and effort is spent on keying the data. OCR engines work well with uniformly sized text. With machine printed data interspersed between pre-printed labels of a different size, OCR engines sometimes read the content erroneously.
 
     
 

STSmartOCR is a suite of image filters that is capable of enhancing images - structured as well as unstructured - making it more OCR friendly and hence improving OCR accuracy. STSmartOCR provides a design interface called STSmartOCRDesigner that allows the developer to select from a wide range of filters - Line Removal, Dash Removal, Comb Removal, Associated Label Removal, Edge Removal, Invert Colors, Erase Between Lines, Block Removal and others.

The designer is a very intuitive application that allows users to configure these filters and test them on sample images before incorporating them into a capture workflow. Some filters have configurable parameters that can be adjusted to obtain the best results.

 

 

 
STSmartOCR
 
     
 

STSmartOCR is known to provide breathtaking results on images where data characters lie in close proximity or on top of the lines or when the labels are too close to the data to be captured. It can also invert colors intelligently so that portions of the image that have white labels on a black background can be reversed to obtain good OCR reads and hence high accuracy results.

 
 
STSmartOCR
 
     
 

STPixID/STEnhanceOCR
With the advent of mobile technology there has been a growing demand for identifying and extracting data from Identification documents such as drivers licenses, passports, national ID cards. The images that are captured range from being color to greyscale to black and white and vary significantly in quality. The font sizes as well as the colors of the fonts and the backgrounds are highly diverse posing a lot of problems for high accuracy OCR capture. Although data to be captured is restricted to a handful of fields, the source documents being unstructured and data sitting on watermarks and patterns in the background poses a challenge. Further the images can be captured in any orientation and can be skewed in any direction.

 
 
STSmartOCR

Using a combination of filters with STPixID and STEnhanceOCR , Stellaris Technologies has been able to carve out a solution that can help in cleaning up the images as well as background, correcting the skews and extracting relevant blocks of data. The solution has been tested against many differnet types of images and has resulted in high OCR accuracy.

 
 

 

 
 
 
 
Company
Products
Expertise
Success Stories

Our Team
Reach Us
Stellaris Technologies Logo
 
Facebook Tweeter linkd Youtube
 
Copyright Reserved 2018 Stellaris Technologies Pvt. Ltd.