IP Library Granted Patent US 10,621,430
Granted Patent B2
US 10,621,430 · App. 15/912,120 · Granted Apr 14, 2020

Determining image forensics using an estimated camera response function

Inventor: Scott McCloskey (Minneapolis, MN)
Assignee: HONEYWELL INTERNATIONAL INC.
G06K9/00577G06K9/6202G06K9/629G06K9/6289G06T1/0028H04N1/32128G06F2221/074G06K2009/00583G06K2209/27G06T2201/0201H04N2201/0005H04N2201/3235H04N2201/3252H04N2201/3257
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 10,621,430
App. No.
15/912,120
Granted
Apr 14, 2020
Kind
B2
Abstract

An image forensics system estimates a camera response function (CRF) associated with a digital image, and compares the estimated CRF to a set of rules and compares the estimated CRF to a known CRF. The known CRF is associated with a make and a model of an image sensing device. The system applies a fusion analysis to results obtained from comparing the estimated CRF to a set of rules and from comparing the estimated CRF to the known CRF, and assesses the integrity of the digital image as a function of the fusion analysis.

Claims (48)

1. A process to assess integrity in a digital image comprising:

estimating a camera response function (CRF) associated with the digital image by computing a plurality of magnitudes of image gradients and modeling distribution in relation to a priori model of natural image gradient magnitudes;

comparing the estimated CRF to a set of rules;

receiving metadata stored in a header of the digital image, the metadata stored in the header of the digital image identifying a make and a model of an image sensing device that created the digital image;

receiving metadata stored in a database, the metadata stored in the database associated with the make and model of the image sensing device that created the digital image;

comparing the metadata stored in the header of the digital image to the metadata stored in the database that is associated with the make and model of the image sensing device that created the digital image;

applying a fusion analysis to results obtained from comparing the estimated CRF to a set of rules and from comparing the metadata stored in the header of the digital image to the metadata stored in the database that is associated with the make and model of the image sensing device that created the digital image; and

assessing the integrity of the image as a function of the fusion analysis.

2. The process of claim 1 , wherein the estimated CRF is estimated using a blurred edge in the digital image.

3. The process of claim 1 , comprising computing image statistics for the digital image, the image statistics to be used with optimization methods to infer the CRF.

4. The process of claim 3 , wherein the image statistics are computed in the spatial domain.

5. The process of claim 3 , wherein the image statistics are computed in a Fourier domain.

6. The process of claim 1 , wherein the set of rules comprises a rule that the CRF is monotonic, a rule that the CRF spans an entire output range of an available bit depth, that the CRF comprises no mid-function peaks along the CRF, that the CRF comprises no negative slopes, that the CRF comprises no mid-function zero slopes, and that a CRF for each color channel of the image sensing device has a substantially similar shape.

7. The process of claim 1 , wherein the fusion analysis comprises one or more of a decision-level fusion, a score-level fusion, and a feature-level fusion.

8. A process to assess integrity in a digital image comprising:

estimating a camera response function (CRF) associated with the digital image by computing a plurality of magnitudes of image gradients and modeling distribution in relation to a priori model of natural image gradient magnitudes;

comparing the estimated CRF to a set of rules;

applying a fusion analysis to results obtained from comparing the estimated CRF to a set of rules; and

assessing the integrity of the digital image as a function of the fusion analysis.

9. The process of claim 8 , comprising:

receiving metadata stored in a header of the digital image, the metadata stored in the header of the digital image identifying a make and a model of an image sensing device that created the digital image;

receiving metadata stored in a database, the metadata stored in the database associated with the make and model of the image sensing device that created the digital image;

comparing the metadata stored in the header of the digital image to the metadata stored in the database that is associated with the make and model of the image sensing device that created the digital image;

applying the fusion analysis to results obtained from comparing the metadata stored in the header of the digital image to the metadata stored in the database that is associated with the make and model of the image sensing device that created the digital image; and

assessing the integrity of the digital image as a function of the fusion analysis.

10. The process of claim 8 , wherein the estimated CRF is estimated using a blurred edge in the digital image.

11. The process of claim 8 , comprising computing image statistics for the digital image, the image statistics to be used with optimization methods to infer the CRF; wherein the image statistics are computed in the spatial domain or the image statistics are computed in the Fourier domain.

12. The process of claim 8 , wherein the set of rules comprises a rule that the CRF is monotonic, a rule that the CRF spans an entire output range of an available bit depth, that the CRF comprises no mid-function peaks along the CRF, that the CRF comprises no negative slopes, that the CRF comprises no mid-function zero slopes, and that the CRF for each color channel of the image sensing device has a substantially similar shape; and wherein the fusion analysis comprises one or more of a decision-level fusion, a score-level fusion, and a feature-level fusion.

13. An image forensics system comprising:

one or more processors; and

a computer readable medium storing instructions that, when executed by the one or more processors, cause the system to perform operations comprising:

estimating a camera response function (CRF) associated with a digital image by computing a plurality of magnitudes of image gradients and modeling distribution in relation to a priori model of natural image gradient magnitudes;

comparing the estimated CRF to a set of rules;

applying a fusion analysis to results obtained from comparing the estimated CRF to a set of rules; and

assessing the integrity of the digital image as a function of the fusion analysis.

14. The image forensics system of claim 13 , comprising:

receiving metadata stored in a header of the digital image, the metadata stored in the header of the digital image identifying a make and a model of an image sensing device that created the digital image;

receiving metadata stored in a database, the metadata stored in the database associated with the make and model of the image sensing device that created the digital image;

comparing the metadata stored in the header of the digital image to the metadata stored in the database that is associated with the make and model of the image sensing device that created the digital image;

applying the fusion analysis to results obtain from comparing the metadata stored in the header of the digital image to the metadata stored in the database that is associated with the make and model of the image sensing device that created the digital image; and

assessing the integrity of the digital image as a function of the fusion analysis.

15. The image forensics system of claim 13 , wherein the estimated CRF is estimated using a blurred edge in the digital image.

16. The image forensics system of claim 13 , comprising computing image statistics for the digital image, the image statistics being computed in the spatial or Fourier domain; and

wherein the estimated CRF is estimated by numerical optimization to explain the difference between observed and expected statistics.

17. The image forensics system of claim 13 , wherein the set of rules comprises a rule that the CRF is monotonic, a rule that the CRF spans an entire output range of an available bit depth, that the CRF comprises no mid-range peaks along the CRF, that the CRF comprises no negative slopes, that the CRF comprises no mid-function zero slopes, and that a CRF for each color channel of the image sensing device has a substantially similar shape.

18. The image forensics system of claim 13 , comprising instructions for computing image statistics for the digital image, the image statistics to be used with optimization methods to infer the CRF.

19. The image forensics system of claim 18 , wherein the image statistics are computed in the spatial domain or the image statistics are computed in a Fourier domain.

20. The image forensics system of claim 13 , wherein the fusion analysis comprises one or more of a decision-level fusion, a score-level fusion, and a feature-level fusion.

Assignments (2)
CONFIRMATORY LICENSE Recorded Mar 6, 2025
From: HONEYWELL INTERNATIONAL INC.
To: GOVERNMENT OF THE UNITED STATES AS REPRESENTED BY THE SECRETARY OF THE AIR FORCE
Reel/Frame 070421/0461 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Oct 5, 2018
From: MCCLOSKEY, SCOTT
To: HONEYWELL INTERNATIONAL INC.
Reel/Frame 047082/0474 →
Continuity (2)
Continuation 15198810 · Jun 30, 2016
Related Publication 20180197004A1 · Jul 12, 2018