Hough processor

• US 10,592,768 B2
• Filed: 08/04/2016
• Issued: 03/17/2020
• Est. Priority Date: 02/04/2014
• Status: Active Grant

• Current Assignee: Fraunhofer Gesellschaft Zur Foerderung Der Angewandten Forsching E.V.
• Original Assignee: Fraunhofer Gesellschaft Zur Foerderung Der Angewandten Forsching E.V.
• Inventors: Krenzer, Daniel, Hess, Albrecht, Katai, Andras
• Primary Examiner(s): Li, Ruiping
• Application Number:
  US15/228,826
• Publication Number:
  US 20160342856A1
• Time in Patent Office: 1,321 Days
• CPC Class Codes: G06F 17/145 Square transforms, e.g. Hadamard, Walsh, Haar, Hough, Slant transforms
• CPC Class Codes: G06T 2207/10012 Stereo images
• CPC Class Codes: G06T 2207/20008 Globally adaptive
• CPC Class Codes: G06T 2207/20061 Hough transform
• CPC Class Codes: G06T 2207/30201 Face
• CPC Class Codes: G06T 3/40 Scaling of whole images or parts thereof, e.g. expanding or contracting
• CPC Class Codes: G06T 3/60 Rotation of whole images or parts thereof
• CPC Class Codes: G06T 5/70 Denoising; Smoothing
• CPC Class Codes: G06T 7/13 Edge detection
• CPC Class Codes: G06T 7/337 involving reference images or patches
• CPC Class Codes: G06T 7/74 involving reference images or patches
• CPC Class Codes: G06T 7/77 using statistical methods
• CPC Class Codes: G06V 10/955 using specific electronic processors
• CPC Class Codes: G06V 40/18 Eye characteristics, e.g. of the iris
• CPC Class Codes: G06V 40/19 Sensors therefor
• CPC Class Codes: G06V 40/193 Preprocessing; Feature extraction
• CPC Class Codes: G06V 40/20 Movements or behaviour, e.g. gesture recognition recognition of facial expressions G06V40/16

Abstract

A Hough processor comprises a pre-processor and a Hough transformation unit. The pre-processor is configured to receive a plurality of samples respectively comprising an image and in order to rotate or reflect the image of the respective sample. The Hough transformation unit is configured to collect a predetermined searched pattern in the plurality of samples on the basis of a plurality of versions. The Hough transformation unit comprises a characteristic being dependent on the searched pattern, which is adjustable according to the searched pattern.

Claims

1. Hough processor comprising the following features:

a pre-processor, which is configured to receive a plurality of samples respectively comprising an image and to rotate the image of the respective sample or to reflect and to output a plurality of versions of the image of the respective sample for each sample; and

a Hough transformation unit, which is configured to collect a predetermined searched pattern within the plurality of samples on the basis of the plurality of versions, wherein a characteristic of the Hough transformation unit, which depends on the searched pattern, is adjustable;

wherein the Hough transformation unit comprises a delay filter the filter characteristic of which depending on the selected searched pattern is adjustable;

wherein the delay filter of the Hough transformation unit comprises one or more delay elements, which are selectively switchable during the ongoing operation in order to allow an adjustment of the filter characteristic of the delay filter;

wherein the Hough transformation unit is connected to a processing unit comprising a unit for analyzing of collected Hough features in order to output a plurality of geometry parameter sets describing the geometry of one or more predefined searched patterns for every sample; and

wherein the processing unit comprises a unit for controlling the adjustable Hough transformation unit in the case of an absent or incorrect recognition of the searched pattern.

2. Hough processor according to claim 1, wherein the Hough transformation unit is configured to determine a pixel-wise correlation.

3. Hough processor according to claim 1, wherein the pre-processor is configured to rotate the image about 360°

/n and to reflect the rotated image, in order to output in parallel n versions of the image per sample, or

wherein the pre-processor is configured to output in series a first set of versions comprising the image rotated about 360°

/n and a second set of versions, comprising the reflected image and the reflected, rotated image, per sample,

wherein n is number of versions.

4. Hough processor according to claim 1, wherein the Hough transformation unit is configured to collect segments of a predetermined pattern, wherein the predetermined patterns originate from a group comprising at least an ellipsis, a circle, a straight line and a combination of straight line and curve.

5. Hough processor according to claim 1, wherein the delay filter comprises per predetermined searched pattern one column with a plurality of delay elements.

6. Hough processor according to claim 1, wherein the Hough transformation unit per delay element of the delay filter comprises a multiplexer, by means of which the respective delay element is selectively connectable in order to change the filter characteristic.

7. Hough processor according to claim 1, wherein the Hough transformation unit per column comprises an amount member, which is configured to add processed signals by means of the one or more delay elements.

8. Hough processor according to claim 7, wherein the Hough transformation unit is configured to collect segments of the predetermined searched pattern and to output a column amount which allows a conclusion to a maximal accordance degree towards one of the predetermined patterns.

9. Hough processor according to claim 1, wherein the Hough transformation unit is configured to output a multi-dimensional Hough room comprising information on the kind of the collected pattern, an accordance degree with the collected pattern, a position of the collected pattern within the image, a size of the collected pattern segments.

10. Hough processor according to claim 9, wherein the Hough transformation unit comprises a filter processor comprising an amount filter and/or a maximum filter for smoothing the position of the identified pattern in the multi-dimensional Hough room.

11. Hough processor according to claim 9, wherein the Hough transformation unit comprises a Hough extractor, which is configured to select one or more searched pattern segments within the Hough room as Hough features, wherein the selection is based on a non-maxima suppression of the Hough room with predetermined thresholds.

12. Hough processor according to claim 1, wherein the unit for controlling the adjustable Hough transformation unit comprises a first mode for adjusting the characteristic during the operation based on the current characteristic and/or based on the current collection results and a second mode for an initial adjustment of the characteristic during the implementation.

13. Hough processor according to claim 2, wherein the correlation is a correlation between a template predetermined by means of the characteristic and an image content.

14. Hough processor according to claim 13, wherein a plurality of templates is predetermined for which a correlation is determined respectively, in order to, thus, acquire an accordance measure with the image content per pixel.

15. Hough processor according to claim 13, wherein every template is assigned to a characteristic of the searched pattern.

16. Hough processor according to claim 13, wherein the Hough transformation unit is configured to output on the basis of the accordance measure per template per pixel a multi-dimensional room, comprising information on the kind of the collected pattern, an accordance measure with the collected pattern, a position of the collected pattern within the image, a size of the collected pattern segments.

17. Hough processor according to claim 1 comprising an image input stage which is arranged between a camera and a pre-processor.

18. Hough processor according to claim 17, wherein the image input stage comprises a unit for collecting of segmentations and/or edges and/or gradients of an image.

19. Hough processor according to claim 1, wherein the Hough processor is implemented in an embedded processor, a programmable logic component, or a client-specific chip.

20. Image analyzing system for tracking a pupil with the following features:

a first Hough path for a first camera, wherein the first Hough path comprises a Hough processor according to claim 1; and

a processing unit comprising a unit for analyzing the collected patterns and for outputting a geometry parameter set, which describes the geometry of one or more predefined searched patterns for each sample.

21. Image analyzing system according to claim 20 the processing unit of which comprises a selective adaptive data processor, which is configured to receive several sets of values, wherein every set is assigned to a respective sample, with the following features:

a filter processor, which is configured to carry out a smoothing of the sets so that based on the received sets, plausible sets are outputted and so that an implausible set is replaced by a plausible one.

22. Image analyzing system according to claim 20 comprising a second Hough path for a second camera of a stereoscopic camera assembly comprising the first and the second camera, wherein the second Hough path comprises a further Hough processor.

23. Image analyzing system according to claim 22, wherein the processing unit comprises a 3D image analyzer, wherein the 3D image analyzer is configured to receive at least one first set of image data, which is determined on the basis of a first image, and a further set of image information, which is determined on the basis of a further image, wherein the first image comprises a pattern of a three-dimensional object from a first perspective in a first image plane and wherein the further set comprises a relative information describing a relative relation between a point of the three-dimensional object and the first image plane, wherein the 3D image analyzer comprises the following features:

a position calculator, which is configured to calculate a position of the pattern in a three-dimensional room based on the first set, a further set, which is determined on the basis of a further image, and a geometric relation between the perspectives of the first and the further image, or in order to calculate the position of the pattern in the three-dimensional room based on the first set and a statistically determined relation between two characteristic features in the first image to each other; and

an alignment calculator configured to calculate a 3D image vector according to which the pattern in the three-dimensional room is aligned, wherein the calculation is based on the first set, the further set and on the calculated position of the pattern.

30. Hough processor according to claim 1, wherein the delay filter is adjusted based on a previous sample of the plurality of samples.

24. Method for Hough processing comprising:

pre-processing of a plurality of samples respectively comprising an image by using a pre-processor, wherein the image of the respective sample is rotated or reflected so that a plurality of versions of the image of the respective sample for each sample is indicated; and

collecting a predetermined pattern in the plurality of samples on the basis of a plurality of versions by using a Hough transformation unit comprising an adjustable characteristic which depends on the searched pattern, wherein the characteristic is adjusted according to the selected set of patterns;

wherein the Hough transformation unit comprises a delay filter the filter characteristic of which depending on the selected searched pattern is adjustable;

wherein the delay filter of the Hough transformation unit comprises one or more delay elements, which are selectively switchable during the ongoing operation in order to allow an adjustment of the filter characteristic of the delay filter; and

wherein the method further comprises the steps of aliasing of connected Hough features in order to output a plurality of geometry parameter sets describing that the geometry of one or more predefined search patterns for a sample and controlling the adjustable Hough transformation unit in the case of an absent of incorrect recognition of the searched pattern.

25. Method according to claim 24, wherein the adjustable characteristic is a filter characteristic of the delay filter.

26. Method according to claim 25, wherein the adjusting of the delay filter is carried out during the implementation or during the ongoing operation, if the pattern is not or incorrectly recognized.

27. Method according to claim 26, wherein the adjustable characteristic is a post-processed characteristic, a curved characteristic, or a distorted characteristic.

28. Method according to claim 24, wherein collecting of the predetermined pattern comprises determining a pixel-wise correlation between a template predetermined by means of a characteristic and an image content, in order to thus acquire an accordance measure with the image content per pixel.

29. A non-transitory digital storage medium having a computer program stored thereon to perform the method for Hough processing comprising:

pre-processing of a plurality of samples respectively comprising an image by using a pre-processor, wherein the image of the respective sample is rotated or reflected so that a plurality of versions of the image of the respective sample for each sample is indicated; and

collecting a predetermined pattern in the plurality of samples on the basis of a plurality of versions by using a Hough transformation unit comprising an adjustable characteristic which depends on the searched pattern, wherein the characteristic is adjusted according to the selected set of patterns;

wherein the Hough transformation unit comprises a delay filter the filter characteristic of which depending on the selected searched pattern is adjustable;

wherein the delay filter of the Hough transformation unit comprises one or more delay elements, which are selectively switchable during the ongoing operation in order to allow an adjustment of the filter characteristic of the delay filter; and

wherein the method further comprises the steps of aliasing of connected Hough features in order to output a plurality of geometry parameter sets describing that the geometry of one or more predefined search patterns for a sample and controlling the adjustable Hough transformation unit in the case of an absent of incorrect recognition of the searched pattern;

when said computer program is run by a computer, an embedded processor, a programmable logic component or a client-specific chip.