Methods and apparatuses for analyzing digital images to automatically select regions of interest thereof

• US 8,126,267 B2
• Filed: 06/27/2007
• Issued: 02/28/2012
• Est. Priority Date: 02/05/2007
• Status: Active Grant

• Current Assignee: Albany Medical College (Albany Medical Center)
• Original Assignee: Albany Medical College (Albany Medical Center)
• Inventors: Talati, Ronak, Barroso, Margarida
• Primary Examiner(s): Patel, Jayesh A
• Application Number:
  US11/769,036
• Publication Number:
  US 20080187241A1
• Time in Patent Office: 1,707 Days
• Field of Search: 382144-145, 382/172, 382/205, 382/272, 382/274, 382/221
• US Class Current: 382/172
• CPC Class Codes: G06V 10/25 Determination of region of interest [ROI] or a volume of interest [VOI]
• CPC Class Codes: G06V 20/693 Acquisition

Abstract

Computer-implemented image processing methods and apparatuses are presented for automatically selecting regions of interest within an image represented by pixel intensity values. A first pixel box is employed in progressively scanning and evaluating the image. If pixels within the first pixel box have pixel-intensity-related characteristics exceeding respective defined thresholds, then those pixels are identified as an area of interest and a second pixel box is employed in progressively scanning and evaluating the selected area of interest to identify regions of interest. Each area of interest is larger than a region of interest, and the second pixel box is smaller than the first. Regions of interest within the image are identified if one or more pixel-intensity-related characteristics of pixels within the second pixel box exceeds a second defined threshold, wherein the second threshold is greater than the first. Once selected, identifying information for the regions of interest is stored or output.

Claims

1. A computer-implemented image processing method comprising:

obtaining an image represented by an array of pixel intensity values;

employing a first pixel box in progressively scanning and evaluating the image for areas of interest, wherein the first pixel box is sized as a first two-dimensional coordinate sub-array of pixels;

selecting one or more areas of interest of the image employing the first pixel box, the selecting comprising identifying an area of interest within the image if at least one pixel-intensity-related characteristic of pixels within the first pixel box at a particular location of the image exceeds a first defined threshold;

employing a second pixel box in progressively scanning and evaluating selected areas of interest of the image for one or more regions of interest, wherein each area of interest is larger than a region of interest and the second pixel box is sized as a second two-dimensional coordinate sub-array of pixels, and wherein the second pixel box is smaller than the first pixel box;

selecting one or more regions of interest of the image employing the second pixel box, the selecting comprising identifying a region of interest within the image if the at least one pixel-intensity-related characteristic of pixels within the second pixel box at a particular location of an area of interest exceeds a second defined threshold, wherein the second defined threshold is greater than the first defined threshold;

performing at least one of storing or outputting location identifying information for selected regions of interest of the image;

wherein employing the second pixel box in progressively scanning and evaluating selected areas of interest further comprises employing a third pixel box in evaluating areas of interest of the image for regions of interest, the third pixel box being sized as a third two-dimensional coordinate sub-array of pixels, wherein the third pixel box is smaller than the second pixel box and shares a common center with the second pixel box, and wherein selecting one or more regions of interest of the image further comprises selecting pixels within the second pixel box at a particular location of an area of interest as a region of interest if a difference between the at least one pixel-intensity-related characteristic of the second pixel box and the at least one pixel-intensity-related characteristic of the third pixel box exceeds a defined threshold; and

wherein the at least one pixel-intensity-related characteristic comprises a mean pixel intensity of pixels, and wherein the second pixel box is selected as a region of interest if the difference between the mean pixel intensity of pixels within the second pixel box at a particular location of an area of interest and the mean pixel intensity of pixels within the third pixel box is greater than a respective defined threshold.

2. The method of claim 1, wherein the second pixel box is an 8×

8 pixel box and the third pixel box is a 6×

6 pixel box.

3. A computer-implemented image processing method comprising;

obtaining an image represented by an array of pixel intensity values;

employing a first pixel box in progressively scanning and evaluating the image for areas of interest, wherein the first pixel box is sized as a first two-dimensional coordinate sub-array of pixels;

selecting one or more areas of interest of the image employing the first pixel box, the selecting comprising identifying an area of interest within the image if at least one pixel-intensity-related characteristic of pixels within the first pixel box at a particular location of the image exceeds a first defined threshold;

employing a second pixel box in progressively scanning and evaluating selected areas of interest of the image for one or more regions of interest, wherein each area of interest is larger than a region of interest and the second pixel box is sized as a second two-dimensional coordinate sub-array of pixels, and wherein the second pixel box is smaller than the first pixel box;

selecting one or more regions of interest of the image employing the second pixel box, the selecting comprising identifying a region of interest within the image if the at least one pixel-intensity-related characteristic of pixels within the second pixel box at a particular location of an area of interest exceeds a second defined threshold, wherein the second defined threshold is greater than the first defined threshold;

performing at least one of storing or outputting location identifyring information for selected regions of interest of the image;

wherein selecting one or more regions of interest of the image employing the second pixel box further comprises identifying a region of interest within the image if multiple pixel-intensity-related characteristics of pixels within the second pixel box at a particular location of the area of interest exceed respective defined thresholds, the multiple pixel-intensity-related characteristics comprising an integrated density of pixels within the second pixel box;

wherein selecting one or more regions of interest within the image employing the second pixel box further comprises identifying a region of interest of the image if pixels within multiple second pixel box locations offset by one pixel in the x-coordinate direction or one pixel in the y-coordinate direction exceed the respective defined thresholds; and

wherein the multiple pixel-intensity-related characteristics further comprise a maximum pixel intensity of pixels within the second pixel box and a mean intensity difference between the second pixel box and a third pixel box, and wherein the mean intensity difference is a difference between the mean intensity of pixels within the second pixel box and the mean intensity of pixels within the third pixel box.

4. A computer-implemented image processing method comprising:

obtaining an image represented by an array of pixel intensity values;

employing a first pixel box in progressively scanning and evaluating the image for areas of interest, wherein the first pixel box is sized as a first two-dimensional coordinate sub-array of pixels;

selecting one or more areas of interest of the image employing the first pixel box, the selecting comprising identifying an area of interest within the image if at least one pixel-intensity-related characteristic of pixels within the first pixel box at a particular location of the image exceeds a first defined threshold;

employing a second pixel box in progressively scanning and evaluating selected areas of interest of the image for one or more regions of interest, wherein each area of interest is larger than a region of interest and the second pixel box is sized as a second two-dimensional coordinate sub-array of pixels, and wherein the second pixel box is smaller than the first pixel box;

selecting one or more regions of interest of the image employing the second pixel box, the selecting comprising identifying a region of interest within the image if the at least one pixel-intensity-related characteristic of pixels within the second pixel box at a particular location of an area of interest exceeds a second defined threshold, wherein the second defined threshold is greater than the first defined threshold;

performing at least one of storing or outputting location identifying information for selected regions of interest of the image; and

further comprising subsequent to selecting regions of interest of the image employing the second pixel box, removing overlapping regions of interest, wherein the removing of overlapping regions of interest further includes selecting from a group of overlapping regions of interest, a region of interest having a highest multiple of (integrated pixel density)×

(mean pixel intensity)×

(mean intensity difference), wherein integrated pixel density is the integrated density of pixels within the respective region of interest, mean pixel intensity is the mean intensity of pixels within the respective region of interest, and mean intensity difference is the difference in mean intensities between pixels of the second pixel box and pixels of a third pixel box employed in selecting the region of interest, wherein the third pixel box is sized as a third two-dimensional coordinate sub-array of pixels which is smaller than the second pixel box and shares a common center with the second pixel box.

5. A computer-implemented image processing method comprising:

obtaining an image represented by an array of pixel intensity values; and

evaluating the image for regions of interest, wherein the evaluating includes;

(i) employing a first pixel box in evaluating the image, the employing including determining for pixels within the first pixel box a pixel-intensity-related characteristic, wherein the first pixel box is sized as a first two-dimensional coordinate sub-array of pixels;

(ii) determining for pixels within a second pixel box the pixel-intensity-related characteristic, wherein the second pixel box is sized as a second two-dimensional coordinate sub-array of pixels, and wherein the second pixel box is smaller than the first pixel box and shares a common center with the first pixel box;

(iii) selecting pixels within the first pixel box as a region of interest of the image if a difference between the pixel-intensity-related characteristic of the first pixel box and the pixel-intensity-related characteristic of the second pixel box exceeds a defined threshold; and

performing at least one of saving or outputting location identifying information for the selected regions of interest of the image;

wherein the selecting further comprises identifying a region of interest if pixels within each of multiple first pixel box locations offset by one pixel in the x-coordinate direction or one pixel in the y-coordinate direction comprise multiple pixel-intensity-related characteristics which exceed respective defined thresholds, the multiple pixel-intensity-related characteristics comprising a mean intensity difference, which is a difference in mean intensities between pixels in the first pixel box at a particular location and the respective second pixel box, a maximum pixel intensity, and an integrated density of pixels within the first pixel box; and

wherein the region of interest is identified if pixels within three first pixel box locations offset by one pixel in the x-coordinate direction or one pixel in the y-coordinate direction respectively comprise a mean intensity difference, maximum pixel intensity and integrated density exceeding their respective defined thresholds.

6. A computer-implemented image processing method comprising:

obtaining an image represented by an array of pixel intensity values and comprising multiple identified regions of interest within the image, wherein at least two regions of interest of the multiple regions of interest within the image overlap; and

selecting one region of interest of the at least two regions of interest overlapping within the image, the selecting comprising;

selecting the region of interest of the at least two regions of interest having a highest multiple of at least two pixel-intensity-related characteristics of pixels within the respective regions of interest of the at least two overlapping regions of interest of the image; and

performing at least one of saving or outputting location identifying information for the selected region of interest having the highest multiple of the at least two pixel-intensity-related characteristics, wherein selecting the region of interest from the at least two regions of interest overlapping further comprises determining a mean intensity difference for each region of interest of the at least two regions of interest, wherein determining the mean intensity difference comprises;

(i) determining for pixels within a first pixel box a mean intensity, wherein the first pixel box is a first two-dimensional coordinate sub-array of pixels sized to the region of interest;

(ii) determining for pixels within a second pixel box the mean pixel intensity, wherein the second pixel box is sized as a second two-dimensional coordinate sub-array of pixels, and wherein the second pixel box is smaller than the first pixel box and shares a common center with the first pixel box; and

(iii) determining a difference between the mean pixel intensity of the first pixel box and the mean pixel intensity of the second pixel box, wherein the mean pixel intensity difference is one pixel-intensity-related characteristic of the at least two pixel-intensity-related characteristics.

7. The method of claim 6, wherein selecting the region of interest from the at least two overlapping regions of interest further comprises selecting the region of interest having the highest multiple of (integrated density)×

(mean intensity)×

(mean intensity difference) for the pixels thereof, wherein integrated pixel density is the integrated density of pixels within the respective region of interest, mean pixel intensity is the mean intensity of pixels within the respective region of interest, and mean intensity difference is the difference in mean intensities between pixels of the first pixel box and pixels of the second pixel box employed in identifying the respective region of interest.