Diagnostic system and method for biological tissue analysis

Patent metadata

Field	Value
Publication number	US-2016232425-A1
Application number	US-201615097780-A
Country	US
Kind code	A1
Filing date	Apr 13, 2016
Priority date	Nov 6, 2013
Publication date	Aug 11, 2016
Grant date	—

Abstract

A computer diagnostic system and related method are disclosed for automatically classifying tissue types in an original tissue image captured by an imaging device based on texture analysis. In one embodiment, the system receives and divides the tissue image into multiple smaller tissue block images. A combination of local binary pattern (LBP), average LBP (ALBP), and block-based LBP (BLBP) feature extractions are performed on each tissue block. The extractions generate a set of LBP, ALBP, and BLBP features for each block which are used to classify its tissue type. The classification results are visually displayed in a digitally enhanced map of the original tissue image. In one embodiment, a tissue type of interest is displayed in the original tissue image. In another or the same embodiment, the map displays each of the different tissue types present in the original tissue image.

First claim

What is claimed is:

1 . A computer-aided diagnostic system for analyzing tissue image data, comprising:
a non-transitory computer readable medium having software program instructions stored thereon; a computer processor communicating with the computer readable medium, the processor when configured with and executing the program instructions being operable to: receive an original tissue image captured by a label free optical imaging device; divide the original tissue image into a plurality of smaller tissue blocks which collectively represent the captured original tissue image, each tissue block having a texture; analyze each tissue block based on its texture to classify a type of tissue found in each tissue block; and generate a digitally enhanced map overlaid on the original tissue image displaying a predetermined tissue type of interest in a visually discernible highlighted manner.

2 . The system according to claim 1 , wherein the processor is further operable to perform on each tissue block a combination of a plurality of texture analysis methods.

3 . The system according to claim 2 , wherein the texture analysis methods comprises local binary pattern extraction that generates a local binary pattern feature, average local binary pattern extraction that generates an average local binary pattern feature, and block based local binary pattern extraction that compares the average intensity value of pixels in blocks of certain shape in a neighborhood around the center pixel.

4 . The system according to claim 3 , wherein the local binary pattern extraction, average local binary pattern extraction and block based local binary pattern extraction for each tissue block is performed in parallel by the processor.

5 . The system according to claim 3 , wherein the processor is further operable to integrate the local binary pattern feature, the average local binary pattern feature and block based local binary pattern feature in each of the tissue blocks and generate a multi-class integrated feature for each tissue block.

6 . The system according to claim 5 , further comprising a neural network classifier implemented by the processor which is configured and operable to classify the type of tissue present in each tissue block based on its respective multi-class integrated feature.

7 . The system according to claim 6 , wherein the neural network classifier performs a tissue pattern recognition analysis on each of the tissue blocks to classify the types of tissues present by correlating the tissue patterns extracted from each tissue block with a predetermined training dataset of tissue patterns.

8 . The system according to claim 5 , wherein the processor is further operable to select attributes from the multi-class integrated feature of each tissue block that represent distinctive characteristics of different types of tissues and eliminate less relevant attributes.

9 . The system according to claim 1 , wherein the original tissue image is in grayscale and tissue of interest is colorized on the digitally enhanced map.

10 . The system according to claim 1 , wherein the imaging device captures the original tissue image by optical coherence microscopy, optical coherence tomography, confocal microscopy, or two photon microscopy.

11 . The system according to claim 1 , further comprising the processor being operable to determine whether each tissue block contains the tissue of interest, wherein if the tissue block contains the tissue of interest a color overlay is displayed on the original tissue image for that tissue block, and if the tissue block does not contain the tissue of interest no color overlay is displayed on the original tissue image for that tissue block.

12 . The system according to claim 10 , wherein the processor is operable to determine a probability that the tissue type found in each tissue block is the tissue of interest, and only displays a color overlay if the probability that the tissue type found is the tissue of interest is larger than the tissue type found being another tissue type.

13 . The system according to claim 1 , wherein the tissue of interest is a tumor.

14 . A computer-aided diagnostic system for analyzing tissue image data, comprising:
a non-transitory computer readable medium having software program instructions stored thereon; a computer processor communicating with the computer readable medium, the processor when configured with and executing the program instructions being operable to: receive an original tissue image captured by a label free optical imaging device; divide the original tissue image into a plurality of smaller tissue blocks which collectively represent the captured original tissue image, each tissue block having a texture; extract a local binary pattern feature, an average local binary pattern feature, and block based local binary pattern features for each tissue block; classify a type of tissue based on the texture features extracted from each tissue block according to pre-defined types of contextually relevant tissues found in the original tissue image; and display a digitally enhanced image of the original tissue image which shows at least one type of tissue in a visually discernible manner.

15 . The system according to claim 14 , wherein the processor is operable to generate a multi-class integrated feature for each tissue block comprising the local binary pattern feature, average local binary pattern feature, and block based local binary pattern features of each tissue block.

16 . The system according to claim 14 , wherein the processor is operable to classify the type of tissue found in each tissue block based on its respective multi-class integrated feature.

17 . The system according to claim 14 , wherein each tissue block in the digitally enhanced image appears in either grayscale or is overlaid with a color to connote a predetermined tissue type of interest.

18 . The system according to claim 14 , wherein the digitally enhanced image comprises multiple tissue types, each tissue type being displayed by the processor with a different visual appearance in the digitally enhanced image.

19 . The system according to claim 17 , wherein each tissue type displayed in the digitally enhanced image is a different color.

20 . A computer aided diagnostic method for automatically classifying tissue types in a digital tissue image, the method implemented by a processor executing program instructions and comprising steps of:
receiving an original tissue image captured by a label free optical imaging device; dividing the original tissue image into a plurality of smaller tissue blocks which collectively represent the captured original tissue image, each tissue block having a texture; classifying a type of tissue found in each tissue block based on its texture; and displaying a digitally enhanced map of the original tissue image in which at least one tissue type is shown in a visually discernible manner from other portions of the original image.

21 . The method according to claim 20 , further comprising a step of the processor extracting a combination of a local binary pattern feature, an average local binary pattern feature, and block based local binary pattern features for each tissue block

22 . The method according to claim 20 , further comprising step of the processor:
integrating the local binary pattern feature, average local binary pattern feature, and block based local binary pattern features for

Assignees

• Univ Lehigh

Inventors

• Huang Xiaolei
• Wan Sunhua
• Zhou Chao

Classifications

• G06V10/806

  of extracted features · CPC title

• G06V10/764

  using classification, e.g. of video objects · CPC title

• G06F18/2431

  Multiple classes · CPC title

• G06T7/0012 Primary

  Biomedical image inspection · CPC title

• G06F18/253

  of extracted features · CPC title