Method for automated radiographic quality assurance

• US 5,734,740 A
• Filed: 10/31/1994
• Issued: 03/31/1998
• Est. Priority Date: 10/31/1994
• Status: Expired due to Fees

• Current Assignee: University of Florida (State University System of Florida)
• Original Assignee: University of Florida (State University System of Florida)
• Inventors: Benn, Douglas K., Shim, Minbo
• Primary Examiner(s): COUSO, YON JUNG
• Application Number:
  US08/332,513
• Time in Patent Office: 1,247 Days
• Field of Search: 382/132, 382/133, 382/168-172, 382/142-144, 382/149, 382/150, 378/207, 364/413.13, 364/413.14
• US Class Current: 382/132
• CPC Class Codes: G06T 7/0012 Biomedical image inspection

Abstract

A method for quantitatively evaluating the quality of radiographic films is disclosed. The method provides for digitizing a radiographic image to form a digital image made up of a plurality of pixels. The individual pixels of the digitized image are assigned grey level values. These grey level values are summed to form a global histogram representing all of the pixels of the digital image. The histogram is subsequently binned and subdivided to form first and second binned histograms. A series of rules is applied to these histograms to evaluate whether the radiographic film is clinically acceptable for use in making a diagnosis.

Claims

1. A computer implemented method for determining film density quality of a clinical radiographic image on an exposed radiographic film comprising the steps of:

exposing a radiographic film to form a radiographic image;

digitizing said radiographic image to form a digital image of said radiographic image, said digital image comprising a plurality of pixels;

storing said digital image in a memory;

assigning a grey level value to each of said plurality of pixels based on a relative darkness of each said pixel;

generating a grey level histogram of said digital image based on said assigned grey level values;

storing said grey level histogram in a memory;

generating at least one binned grey level histogram of said digital image based on said assigned grey level values;

storing said at least one binned grey level histogram in a memory; and

determining a quality classification of said radiographic image based on an evaluation of said binned grey level histogram of said digital image;

wherein the step of generating said at least one binned grey level histogram comprises

retrieving said grey level histogram from memory;

summing a number of pixels of said digital image having grey level values in a first range, said first range being defined as a dark range;

summing a number of pixels of said digital image having grey level values in a second range, said second range being defined as an intermediate range;

summing a number of pixels of said digital image having grey level values in a third range, said third range being defined as a light range;

said first binned histogram reflecting relative percentages of the numbers of pixels of the digital images in each of said first, second and third ranges; and

storing said first binned histogram in a memory.

2. The method of claim 1, wherein said grey level values are in the range of 0 to 255, 0 being a darkest grey level and 255 being a lightest grey level.

3. The method of claim 1, comprising the further steps of:

retrieving said digital image from memory;

dividing said digital image into three regions, each region being approximately one-third of the digital image and each of the three regions being independent of each other;

assigning grey level values to each of said plurality of pixels;

generating a grey level histogram for each of said three independent regions; and

storing said grey level histogram of each said three independent regions in a memory.

4. The method of claim 3, further comprising the step of binning the grey level histograms of each of the three regions to form a second binned grey level histogram comprising a binned grey level histogram for each said region, wherein the step of binning comprises:

retrieving said grey level histogram for each said three independent regions from said memory;

summing a number of pixels of a given region of said digital image having grey level values in said first range;

summing a number of pixels in a given region of said digital image having grey level values in said second range;

summing a number of pixels in a given region of said digital image having grey level values in said third range;

repeating said summing steps for each of said three independent regions;

generating a second binned histogram for each said region, wherein said binned histogram reflects relative percentages of said first, second and third ranges in each region; and

storing said second binned histogram in a memory.

5. The method of claim 4 wherein said step of determining quality classification of said radiographic image comprises:

retrieving said first and second binned histograms from memory;

comparing the relative percentage of said first, second and third ranges in said first binned histogram and comparing the relative percentages of said first, second and third ranges in each of the three regions of said second binned histogram in accordance with predetermined rules, said rules comprising at least one of a first rule, a complex dark rule and a complex light rule; and

determining the quality classification based on a result of applying said predetermined rules to said binned histogram.

6. The method of claim 5, wherein applying said rules comprises:

applying said first rule to said first binned histogram, said first rule comprising;

classifying the quality of said radiographic image as acceptable when a percentage of pixels in said first range is less than 65% and a percentage of pixels in said third range is less than 65%;

applying said complex dark rule to said second binned histogram to determine if the quality of said radiographic image is acceptable or too dark when a percentage of pixels of said first binned histogram in said first range is greater than 65% in each region, said complex dark rule comprising;

classifying the quality of said radiographic image as acceptable if the percentage of pixels in the first range is less than 65% in any of the three regions;

classifying the quality of the radiographic image as too dark if the percentage of pixels in the first range is not less than 65% in each of the three regions; and

applying said complex light rule to said second binned histogram to determine if the quality of said radiographic image is acceptable or too light when the percentage of pixels in the third range of said first binned histogram is greater than 65%, said complex light rule comprising;

classifying the quality of the radiographic image as too light when the percentage of pixels in the third range is greater than 65% in all three regions;

classifying the quality of the radiographic image as too light when the percentage of pixels in the third range is greater than 70% in two of the three regions and greater than 50% in a remaining region; and

classifying the quality of the radiographic image as acceptable if it has not been determined to be too light.

7. The method of claim 1, wherein:

said first range is defined as a range of grey level values from 0 to 85;

said second range is defined as a range of grey level values from 86 to 170; and

said third range is defined as a range of grey level values from 171 to 255.