Browse Prior Art Database

Iterated active contour for text segmentation

IP.com Disclosure Number: IPCOM000200577D
Publication Date: 2010-Oct-19
Document File: 5 page(s) / 312K

Publishing Venue

The IP.com Prior Art Database

Abstract

Please enter 2-3 line Abstract

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 37% of the total text.

Page 01 of 5

There is still difficulty in extracting textual data from images acquired from different kinds of media,

whether business cards or posters hanging at a bus station.

Most of

the current OCR applications decode images that are captured by well-calibrated flat bed scanners, but fail to handle noisy text.

    The emergence of mobile phones equipped with high resolution cameras, audio recording facilities, memory, and processing capabilities makes them ideally suited for acquiring information in real time. Example of such information is encountered while we are standing in a bus station we are surrounded with a lot of posters and advertisements.

Another example,

                known as cross talk (or show-through) interference, is a common occurrence when scanning duplex printed documents. The backside printing shows through the paper thus contaminating the front side image.

    This disclosure introduces a novel method to extract text from complicated images. The method is robust in handling text that interferes with other text or graphics.

    We present a new method to detect extract text from images. The process included the following steps:
1. Find global kernels of text in image (optionally).
2.

Apply active contour on the global kernels of text or on the entire image

Apply iteratively active contour on the resulted masks.

4. Stop iterations when masks are of uniform text (intensity/color)
5. Local masks refinement


In general, active contour splits space into two regions, under several

restrictions. In case we have text in image in several intensities or colors, the above iterated path is resulted with masks for every type of text. Parts of the resulted masks are artifacts; e.g. the show through effects discussed above.

1. Background: What is the problem solved by your invention? Describe known solutions to this problem (if any). What are the drawbacks of such known solutions, or why is an additional solution required? Cite any relevant technical documents or references.

There is still difficulty in extracting textual data from images acquired from different kinds of media,

whether business cards or posters hanging at a bus station.

Most of

the current OCR applications decode images that are captured by well-calibrated flat bed scanners, but fail to handle noisy text.

The emergence of mobile phones equipped with high resolution cameras, audio recording facilities, memory, and processing capabilities makes them ideally suited for acquiring information in real time. Example of such information is encountered

while we are standing in a bus station we are surrounded with a lot of posters and

advertisements.

Another example,

            known as cross talk (or show-through) interference, is a common occurrence when scanning duplex printed documents. The backside printing shows through the paper thus contaminating the front side image.

This disclosure introduces a novel method to extract text from complicated images. The method is robust in handling text that interferes with...