System and method for diagnosis of bovine diseases using auscultation analysis

• US 8,152,734 B2
• Filed: 11/07/2008
• Issued: 04/10/2012
• Est. Priority Date: 11/28/2007
• Status: Active Grant

• Current Assignee: Intervet Incorporated
• Original Assignee: Pierson Precision Auscultation
• Inventors: Noffsinger, Thomas H., Taylor, Garrett W., Taylor, Wade A.
• Primary Examiner(s): Foley, Shanon A
• Application Number:
  US12/267,448
• Publication Number:
  US 20090137918A1
• Time in Patent Office: 1,250 Days
• Field of Search: 600/529
• US Class Current: 600/529
• CPC Class Codes: A61B 2503/40 Animals
• CPC Class Codes: A61B 5/08 Detecting, measuring or recording devices for evaluating the respiratory organs A61B5/0205 takes precedence
• CPC Class Codes: A61B 5/7257 using Fourier transforms
• CPC Class Codes: A61B 7/003 Detecting lung or respiration noise
• CPC Class Codes: A61B 7/04 Electric stethoscopes microphones, acoustic transducers therefor H04R
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems

Abstract

A system and method are provided for diagnosis of bovine respiratory diseases using auscultation techniques. Acoustic characteristics of a recorded spectrogram are compared with existing data enabling a diagnosis to be made for a diseased animal. Lung sounds are obtained by use of an electronic stethoscope, and the sounds are stored as digital data. Signal conditioning is used to place the data in a desired format and to remove undesirable noise associated with the recorded sounds. An algorithm is applied to data, and lung scores are calculated. The lung scores are then categorized into various levels of perceived pathology based upon baseline data that categorizes the lung scores. From the lung scores, a caregiver can associate a diagnosis, prognosis, and a recommended treatment. Analysis software generates the lung scores from the recorded sounds, and may also provide a visual display of presumptive diagnoses as well as recommended treatments.

Claims

1. A system for diagnosing bovine diseases using auscultation analysis, said system comprising:

a processor for processing auscultated lung sounds obtained from a bovine in the form of digital sound data detected by a stethoscope;

computer coded instructions for manipulating the digital data through incorporation of at least one algorithm used to calculate a numerical lung score, said algorithm utilizing values of selected frequencies of the auscultated sounds;

a database for storing data reflective of diagnoses, treatments, and prognoses that correspond to a plurality of baseline numerical lung scores; and

a user interface for displaying a spectrogram reflective of the auscultated lung sounds, and displaying the lung score as it is associated with at least one of a corresponding diagnosis, treatment, and prognosis.

2. A system, as claimed in claim 1, wherein:

said stethoscope is an electronic stethoscope that detects auscultated sounds, and converts the auscultated sounds to a digital format for transfer to the processor as the digital sound data.

3. A system, as claimed in claim 1, wherein:

the detected sound is loaded into the computer from the stethoscope as a .wav file.

4. A system, as claimed in claim 1, wherein:

the digital sound data is stored in an array, and a short-time Fourier transform is performed on the digital sound data to place the sound data in a frequency domain.

5. A system, as claimed in claim 4, wherein:

the data from said Fourier transform is separated into groups of amplitudes of frequencies as follows;

500-540 Hz, 540-580 Hz, 580-620 Hz, 620-660 Hz, 660-700 Hz, 700-740 Hz, 740-780 Hz, 780-820 Hz, 820-860 Hz, 860-900 Hz.

6. A system, as claimed in claim 5, wherein:

an algorithm is applied to the data in said groups of amplitudes of frequencies to obtain a calculated lung score, said algorithm being expressed by the following;

x₁=result of 500-540 Hz FIR filter;

x₂=result of 540-580 Hz FIR filter;

x₃=result of 580-620 Hz FIR filter;

x₄=result of 620-660 Hz FIR filter;

x₅=result of 660-700 Hz FIR filter;

x₆=result of 700-740 Hz FIR filter;

x₆=result of 700-740 Hz FIR filter;

x₇=result of 740-780 Hz FIR filter;

x₈=result of 780-820 Hz FIR filter;

x₉=result of 820-860 Hz FIR filter;

x₁₀=result of 860-900 Hz FIR filter;

score=0.205x₁+0.075x₂+0.02x₃+0.2x₄+0.35x₅+0.02x₆+0.02x₇+0.09x₈+0.01x₉+0.01x₁₀.

7. A system, as claimed in claim 1, wherein:

said lung score is compared to baseline data in the form of threshold values that generally correspond to bovine respiratory conditions, and wherein a recommended treatment is generated.

8. A system, as claimed in claim 1, further including:

means for removing selected frequencies from said spectrogram attributed to noise whereby the lung score is calculated taking into account the removed frequencies.

9. A system, as claimed in claim 1, wherein:

wherein said lung score displayed is at least one of a calculated lung score or a scaled lung score.

10. A system, as claimed in claim 7, wherein:

said threshold values correspond to said conditions as follows;

a) condition is low normal when said calculated lung score is between about 0 and 74.5

b) condition is high normal when said calculated lung score is between about 74.5 and 149

c) condition is low mild acute when said calculated lung score is between about 150 and 165

d) condition is high mild acute when said calculated lung score is between about 165 and 180

e) condition is low severe acute when said calculated lung score is between about 181 and 250.5

f) condition is high severe acute when said calculated lung score is between about 250.5 and 319

g) condition is low chronic when said calculated lung score is between about 320 and 400

h) condition is median chronic when said calculated lung score is between about 400 and 500; and

i) condition is high chronic when said calculated lung score is greater than 500.

11. A method for diagnosing bovine diseases using auscultation analysis, said method comprising:

recording auscultated sounds emitted from a bovine and converting the sounds to digital data;

applying a short-time Fourier transform on the digital data to convert the digital data to data in a frequency domain;

determining amplitudes of frequencies present in the converted data in a range between about 500 and 900 Hz;

separating the converted data having the amplitudes within the 500 to 900 Hz range into pre-determined groups;

applying an algorithm to the converted data in the pre-determined groups to generate a lung score;

comparing the lung score to baseline data, said baseline data indicating a level of pathology within the bovine based on the magnitude of the lung score;

making at least one of a diagnosis, prognosis, and treatment recommendation based upon said comparison; and

displaying the lung score and at least one of the diagnosis, prognosis or recommended treatment on a user interface.

12. A method, as claimed in claim 11, wherein:

the data from said Fourier transform is separated into groups of amplitudes of frequencies as follows;

500-540 Hz, 540-580 Hz, 580-620 Hz, 620-660 Hz, 660-700 Hz, 700-740 Hz, 740-780 Hz, 780-820 Hz, 820-860 Hz, 860-900 Hz.

13. A method, as claimed in claim 12, wherein:

an algorithm is applied to the data in said groups of amplitudes of frequencies to obtain a calculated lung score, said algorithm being expressed by the following;

x₁=result of 500-540 Hz FIR filter;

x₂=result of 540-580 Hz FIR filter;

x₃=result of 580-620 Hz FIR filter;

x₄=result of 620-660 Hz FIR filter;

x₅=result of 660-700 Hz FIR filter;

x₆=result of 700-740 Hz FIR filter;

x₆=result of 700-740 Hz FIR filter;

x₇=result of 740-780 Hz FIR filter;

x₈=result of 780-820 Hz FIR filter;

x₉=result of 820-860 Hz FIR filter;

x₁₀=result of 860-900 Hz FIR filter;

score=0.205x₁+0.075x₂+0.02x₃+0.2x₄+0.35x₅+0.02x₆+0.02x₇+0.09x₈+0.01x₉+0.01x₁₀.

14. A method, as claimed in claim 11, wherein:

said lung scores compared to baseline data in the form of threshold values that generally correspond to bovine respiratory conditions, wherein a recommended treatment is generated; and

said threshold values correspond to said conditions as follows;

a) condition is low normal when said calculated lung score is between about 0 and 74.5

b) condition is high normal when said calculated lung score is between about 74.5 and 149

c) condition is low mild acute when said calculated lung score is between about 150 and 165

d) condition is high mild acute when said calculated lung score is between about 165 and 180

e) condition is low severe acute when said calculated lung score is between about 181 and 250.5

f) condition is high severe acute when said calculated lung score is between about 250.5 and 319

g) condition is low chronic when said calculated lung score is between about 320 and 400

h) condition is median chronic when said calculated lung score is between about 400 and 500

i) condition is high chronic when said calculated lung score is greater than 500.

15. A method for diagnosing bovine diseases using auscultation analysis, said method comprising:

recording auscultated sounds emitted from a bovine and converting the sounds to digital data;

converting the digital data to data in a frequency domain;

determining amplitudes of frequencies present in the converted data in a range between about 500-900 Hz;

separating the converted data having the amplitudes within the 500-900 Hz range into predetermined groups;

applying an algorithm to the converted data in the predetermined group to generate a lung score;

evaluating the temperature of the animal and a projected market date of the animal;

comparing the lung score to baseline data, said baseline data indicating a level of pathology within the bovine based on the magnitude of the lung score;

making at least one of a diagnosis, prognosis, and treatment recommendation based upon said comparison; and

displaying the lung score and at least one of the diagnosis, prognosis, or recommended treatment on a user interface.

16. A method, as claimed in claim 15, wherein:

said recommended treatment includes administration of a product in a preselected category.

17. A method, as claimed in claim 15, wherein:

converting the digital data to data in the frequency domain is conducted by applying a short-time Fourier transform on the digital data.

18. A method, as claimed in claim 15, wherein:

said lung score is compared to baseline data in the form of threshold values that generally correspond to bovine respiratory conditions, and wherein a recommended treatment is generated;

said threshold values corresponding to said conditions as follows;

a) condition is low normal when said calculated lung score is between about 0 and 74.5

b) condition is high normal when said calculated lung score is between about 74.5 and 149

c) condition is low mild acute when said calculated lung score is between about 150 and 165

d) condition is high mild acute when said calculated lung score is between about 165 and 180

e) condition is low severe acute when said calculated lung score is between about 181 and 250.5

f) condition is high severe acute when said calculated lung score is between about 250.5 and 319

g) condition is low chronic when said calculated lung score is between about 320 and 400

h) condition is median chronic when said calculated lung score is between about 400 and 500

i) condition is high chronic when said calculated lung score is greater than 500.

19. A method, as claimed in claim 15, wherein:

said algorithm is expressed by the following;

x₁=result of 500-540 Hz FIR filter;

x₂=result of 540-580 Hz FIR filter;

x₃=result of 580-620 Hz FIR filter;

x₄=result of 620-660 Hz FIR filter;

x₅=result of 660-700 Hz FIR filter;

x₆=result of 700-740 Hz FIR filter;

x₆=result of 700-740 Hz FIR filter;

x₇=result of 740-780 Hz FIR filter;

x₈=result of 780-820 Hz FIR filter;

x₉=result of 820-860 Hz FIR filter;

x₁₀=result of 860-900 Hz FIR filter;

score=0.205x₁+0.075x₂+0.02x₃+0.2x₄+0.35x₅0.02x₆+0.02x₇+0.09x₈+0.01x₉+0.01x₁₀.