Method for identifying implanted reconstructive prosthetic devices

• US 8,249,318 B2
• Filed: 09/26/2008
• Issued: 08/21/2012
• Est. Priority Date: 09/26/2008
• Status: Active Grant

• Current Assignee: Osteoware, Inc.
• Original Assignee: Osteoware, Inc.
• Inventors: Schmitt, Mike, Bergquist, Troy S., Smithgall, Brian
• Primary Examiner(s): Le, Vu
• Assistant Examiner(s): Cese, Kenny
• Application Number:
  US12/239,531
• Publication Number:
  US 20100080426A1
• Time in Patent Office: 1,425 Days
• Field of Search: 382/128, 382/131, 382/132
• US Class Current: 382/128
• CPC Class Codes: G06F 18/24147 Distances to closest patterns, e.g. nearest neighbour classification
• CPC Class Codes: G06F 18/253 of extracted features
• CPC Class Codes: G06F 18/40 Software arrangements specially adapted for pattern recognition, e.g. user interfaces or toolboxes therefor
• CPC Class Codes: G06V 10/764 using classification, e.g. of video objects
• CPC Class Codes: G06V 10/945 User interactive design; Environments; Toolboxes
• CPC Class Codes: G06V 2201/033 of skeletal patterns

Abstract

A method for identifying implanted reconstructive prosthetic devices comprising obtaining a digital radiographic image of a prosthetic implant that has been implanted in a person or animal for which the manufacturer and/or model is unknown; allowing a user to enter into a computer metadata relating to the implant for use as metadata feature values; cleaning up the unknown implant image; rotating, flipping and/or scaling the unknown implant image; extracting features from the unknown implant image according to one or more feature extraction algorithms; and comparing the metadata and extracted feature values for the unknown implant image to feature values for other implant images according to a comparison algorithm to create an overall likelihood score for each of the other implant images.

Claims

1. A method for identifying implanted reconstructive prosthetic devices comprising:

(a) obtaining a digital radiographic image of a prosthetic implant that has been implanted in a person or animal for which the manufacturer and/or model is unknown;

(b) allowing a user to enter into a computer metadata relating to identification of the unknown implant for use as metadata feature values;

(c) cleaning up the unknown implant image by removing extraneous image content not relevant to identification of the unknown implant, converting the unknown implant image to monochrome, thresholding it, and/or smoothing out any rough edges;

(d) providing a database of images of implants for which the manufacturer and/or model is known, wherein each of the known images has a format;

(e) rotating, flipping and/or scaling the unknown implant image to a normalized format that matches the format of the images in the database of known implants;

(f) extracting features from the unknown implant image according to one or more feature extraction algorithms to generate extracted feature values, wherein the unknown implant image has a shape, the extracted feature values describe the shape of the unknown implant image, the metadata feature values and the extracted feature values for the unknown implant image are expressed as a feature vector, and the known images stored in the database also have feature vectors;

(g) comparing the feature vector for the unknown implant image to the feature vectors of the known images stored in the database according to a comparison algorithm to create an overall likelihood score for each of the known implant images; and

(h) wherein step (g) generates results, displaying the results to a user, wherein the results that are displayed to the user include feature match strength descriptors that reflect likelihood of a match between a feature of the unknown implant image and a feature of one of the known implant images in the database.

3. The method of claim 1 or 2, wherein the comparison algorithm is:

4. The method of claim 1 or 2, wherein the overall likelihood score is displayed to the user as an overall match strength descriptor.

5. The method of claim 1, wherein the metadata comprises one or more of the following:

date of implantation;

implant manufacturer;

implanting surgeon name;

hospital of implantation;

type of surgical procedure; and

descriptors that describe a physical aspect of the implant.

6. The method of claim 1, further comprising:

(h) providing the user with an option to include the unknown implant image and associated metadata and extracted features in the database.

8. The method of claim 1 or 2, wherein the results displayed to the user include one or more of the following:

job information;

raw and/or processed radiographic implant images;

annotated implant images;

feature values;

feature match strength descriptors;

overall likelihood scores; and

overall match strength descriptors.

9. The method of claim 1 or 2, wherein one of the extracted features is a match pattern, and wherein a pattern match quality score reflects likelihood of a match between a match pattern of the unknown implant image and a match pattern of one of the known implant images in the database.

2. A method for identifying implanted reconstructive prosthetic devices comprising:

(a) obtaining a digital radiographic image of an prosthetic implant that has been implanted in a person or animal for which the manufacturer and/or model is unknown;

(b) cleaning up the unknown implant image by removing extraneous image content not relevant to identification of the unknown implant, converting the unknown implant image to monochrome, thresholding it, and/or smoothing out any rough edges;

(c) providing a database of images of implants for which the manufacturer and/or model is known, wherein each of the known images has a format;

(d) rotating, flipping and/or scaling the unknown implant image to a normalized format that matches the format of the images in the database of known implants;

(e) extracting features from the unknown implant image according to one or more feature extraction algorithms to generate extracted feature values, wherein the unknown implant image has a shape, the extracted feature values describe the shape of the unknown implant image, the extracted feature values for the unknown implant image are expressed as a feature vector, and the known images stored in the database also have feature vectors;

(f) comparing the feature vector for the unknown implant image to the feature vectors of the known images stored in the database according to a comparison algorithm to create an overall likelihood score for each of the known implant images; and

(g) wherein step (f) generates results, displaying the results to a user, wherein the results that are displayed to the user include feature match strength descriptors that reflect likelihood of a match between a feature of the unknown implant image and a feature of one of the known implant images in the database.

7. The method of claim 2, further comprising:

(g) providing the user with an option to include the unknown implant image and associated extracted features in the database.