×

Identifying regions of text to merge in a natural image or video frame

  • US 9,053,361 B2
  • Filed: 01/23/2013
  • Issued: 06/09/2015
  • Est. Priority Date: 01/26/2012
  • Status: Expired due to Fees
First Claim
Patent Images

1. A method of identifying regions to merge, in an image of a scene of real world captured by a camera in a handheld device, the method of identifying regions comprising:

  • checking whether a first block, which contains a first region of pixels that are contiguous with one another and comprising a local extrema of intensity in the image, satisfies a predetermined test, for presence along a line, of pixels with intensities binarizable to a common value;

    marking the first block as pixel-line-present, in a memory, when a result of the checking indicates the predetermined test is satisfied;

    identifying a second block that is located in the image adjacent to the first block, wherein at least the first block is marked as pixel-line-present;

    merging a first set of positions indicative of the first region of pixels in the first block with a second set of positions indicative of a second region of pixels in the second block to obtain a merged set of positions in a merged block, when a predetermined rule is satisfied by one or more geometric attributes of the first block and the second block;

    wherein the first region of pixels and the second region of pixels do not contact one another in the merged block;

    wherein the merging is performed prior to classification of any pixel in the first region of pixels and in the second region of pixels as text or non-text; and

    re-doing the checking, on the merged block, to determine whether the merged block satisfies the predetermined test;

    wherein one or more of the checking, the marking, the identifying, the merging, and the re-doing are performed by at least one processor coupled to the memory.

View all claims
  • 1 Assignment
Timeline View
Assignment View
    ×
    ×