Endoscopic image enhancement using contrast limited adaptive histogram equalization (CLAHE) implemented in a processor

Patent metadata

Field	Value
Publication number	US-10791308-B2
Application number	US-201916685299-A
Country	US
Kind code	B2
Filing date	Nov 15, 2019
Priority date	May 17, 2015
Publication date	Sep 29, 2020
Grant date	Sep 29, 2020

Abstract

Systems and methods of enhancing images use a contrast limited adaptive histogram equalization (CLAHE) algorithm in a field programmable gate array (FPGA). The images may be obtained by the imaging elements of a multiple imaging elements endoscope of an endoscopy system.

First claim

We claim:

1. An endoscope system comprising:
an endoscope including a first viewing element; and
a control unit operatively coupled to the endoscope by a cable, wherein the control unit comprises:
a base board module;
one or more processing boards that implement a contrast limited adaptive histogram equalization (CLAHE) algorithm to enhance images obtained by the first viewing element, wherein the one or more processing boards perform the steps of:
receiving a first video stream from the first viewing element,
building first histogram information based on a first portion of the first video stream,
building second histogram information based on a second portion of the first video stream,
converting the first histogram information into first cumulative distribution function values including applying a first clipping threshold to the first histogram information,
converting the second histogram information into second cumulative distribution function values including applying a second clipping threshold to the second histogram information, wherein the second clipping threshold is different from the first clipping threshold,
obtaining a first plurality of frames from the first video stream,
rearranging the first plurality of frames to generate a plurality of first delayed video streams,
generating a first enhanced video stream based on the plurality of first delayed video streams, the first cumulative distribution function values, and the second cumulative distribution function values, and
sending an output video stream indicative of the first enhanced video stream to a display.

2. The endoscope system of claim 1 , wherein the first video stream is received by an adapter of the one or more processing boards, the adapter having a conversion module, and wherein the first video stream is converted from a first color space video stream to a second color space video stream by the conversion module.

3. The endoscope system of claim 1 , wherein the first plurality of frames are obtained from the first video stream using a frame grabber of the one or more processing boards.

4. The endoscope system of claim 3 , wherein the base board module further comprises a random access memory module in communication with the frame grabber.

5. The endoscope system of claim 1 , wherein the first plurality of frames are rearranged to generate the first plurality of delayed video streams using a fetching module of the one or more processing boards.

6. The endoscope system of claim 1 , wherein the histogram information is converted into cumulative distribution function values, and the first enhanced video stream is generated using an interpolation algorithm module of the one or more processing boards.

7. The endoscope system of claim 1 , wherein building first histogram information based on the first video stream includes splitting the first video stream into a first plurality of tiles, wherein units of the tile heights are in video lines and units of the tile widths are in pixels, wherein each of the first plurality of tiles overlaps with its neighboring tiles.

8. A method of enhancing images obtained by an endoscope system using a contrast limited adaptive histogram equalization (CLAHE) process, wherein said endoscope system comprises (a) one or more processing boards configured to implement said process, and (b) an endoscope operatively coupled to the one or more processing boards, the endoscope including a first viewing element, said method comprising the steps of:
receiving a first video stream from the first viewing element;
splitting the first video stream into a plurality of tiles, wherein units of the tile heights are in video lines and units of the tile widths are in pixels;
building histogram information based on the plurality of tiles;
converting the histogram information into cumulative distribution function values;
obtaining the first video stream as a first plurality of frames;
rearranging the first plurality of frames to generate a first delayed video stream;
generating an enhanced video stream, using the first delayed video stream and the cumulative distribution function values; and
displaying an output video stream indicative of the enhanced video stream.

9. The method of claim 8 , wherein each of the plurality of tiles overlaps with its neighboring tiles.

10. The method of claim 8 , further comprising converting the first video stream from a first color space video stream to a second color space video stream, including the steps of:
performing YCbCr to RGB conversion using MAD instantiation;
performing a function similar to Y(x)=x{circumflex over ( )}02.40 using a LUT;
performing RGB to XYZ conversion using MAD instantiation;
executing a function similar to Y(x)=x{circumflex over ( )}00.33 using a LUT; and
performing XYZ to Lab conversion using MAD instantiation.

11. The method of claim 8 , wherein the enhanced video stream is an enhanced Lab color space video stream, and further comprising converting the enhanced Lab color space video stream to an enhanced YCbCr color space video stream.

12. The method of claim 11 , wherein converting the enhanced Lab color space video stream to the enhanced YCbCr color space video stream comprises the steps of:
performing Lab to XYZ conversion using MAD instantiation;
executing a function similar to Y(x)=x{circumflex over ( )}00.33 using a LUT;
performing XYZ to RGB conversion using MAD instantiation;
performing a function similar to Y(x)=x{circumflex over ( )}(1/2.40) using a LUT; and
performing RGB to YCbCr conversion using MAD instantiation.

13. A non-transient computer readable medium containing program instructions for causing a computer to perform a method of enhancing images obtained by a first viewing element in an endoscope system using a contrast limited adaptive histogram equalization (CLAHE) process, wherein said endoscope system comprises a processor configured to implement said process, and an endoscope operatively coupled to the processor, the endoscope including the first viewing element, said method comprising the steps of:
receiving a first video stream from the first viewing element,
building first histogram information based on a first component of the first video stream,
building second histogram information based on a second component of the firs video stream,
converting the first histogram information into first cumulative distribution function values including applying a first clipping threshold to the first histogram information,
converting the second histogram information into second cumulative distribution function values including applying a second clipping threshold to the second histogram information, wherein the second clipping threshold is different from the first clipping threshold,
obtaining a first plurality of frames from the first video stream,
rearranging the first plurality of frames to generate a plurality of first delayed video streams,
generating a first enhanced video stream based on the plurality of first delayed video streams, the first cumulative distribution function values, and the second cumulative distribution function values, and
sending an output video stream indicative of the first enhanced video stream to a display.

14. The computer readable medium of claim 13 , the method further comprising the step of converting the first video stream from a first color space video stream to a second color space video stream, comprising the steps of:
performing YCbCr to RGB conversion using MAD instantiation;
performing a function similar to Y(x)=x{circumflex over ( )}2.40 using a LUT;
performing RGB to XYZ conversio

Assignees

• Endochoice Inc

Inventors

• Sidar Itay
• Davidson Tal
• Kronman Achia
• Mor Lior
• Levy Idan

Classifications

• A61B1/000095

  for image enhancement · CPC title

• G06T2200/28

  involving image processing hardware · CPC title

• G06T5/20

  using local operators · CPC title

• A61B1/0005

  combining images e.g. side-by-side, superimposed or tiled · CPC title

• G06T5/40

  using histogram techniques · CPC title