AU2018100159A4 - Method of sorting objects in an image - Google Patents
Method of sorting objects in an image Download PDFInfo
- Publication number
- AU2018100159A4 AU2018100159A4 AU2018100159A AU2018100159A AU2018100159A4 AU 2018100159 A4 AU2018100159 A4 AU 2018100159A4 AU 2018100159 A AU2018100159 A AU 2018100159A AU 2018100159 A AU2018100159 A AU 2018100159A AU 2018100159 A4 AU2018100159 A4 AU 2018100159A4
- Authority
- AU
- Australia
- Prior art keywords
- image
- objects
- category
- computer system
- feb
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000001914 filtration Methods 0.000 claims description 3
- 239000000284 extract Substances 0.000 abstract description 3
- 239000011435 rock Substances 0.000 description 26
- 230000015654 memory Effects 0.000 description 12
- 238000003860 storage Methods 0.000 description 7
- 230000001788 irregular Effects 0.000 description 4
- 239000003795 chemical substances by application Substances 0.000 description 3
- 239000003086 colorant Substances 0.000 description 3
- 238000009826 distribution Methods 0.000 description 3
- 239000004575 stone Substances 0.000 description 3
- 238000007796 conventional method Methods 0.000 description 2
- 238000013500 data storage Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 239000000654 additive Substances 0.000 description 1
- 230000000996 additive effect Effects 0.000 description 1
- 238000007621 cluster analysis Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000005553 drilling Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
Landscapes
- Image Analysis (AREA)
Abstract
A computer system executes a method that improves the accuracy of sorting of objects in an image. The method extracts objects of a category in an image and reduces noise in the category. A second image is generated from the image with the category of objects in white color and the background in black color. A marked object is constructed in the second image by a pixel with at least two non-zero neighbor pixels. The number of the marked objects in the second image is counted. Defining a category in an image Extracting objects of the category Reducing errors in the category Generating a second image with the category of objects in white color and the background in black color Improving accuracy of sorting objects in the image Constructing a marked object by a pixel with at least two non-zero neighbor Counting a number of the marked objects in the binary image Fig.1
Description
FIELD OF THE INVENTION
The present invention relates to a method that sorts objects in an image.
BACKGROUND
Analyzing images requires categorization and statistics of numerous irregular shaped objects. The characteristics information, such as the number of interested objects of each category, average size and size distribution of the objects of each category, etc. need to be extracted.
New method that sorts objects in an image will meet advancing technological needs and assist in advancing this technology field.
2018100159 05 Feb 2018
SUMMARY OF THE INVENTION
One example embodiment is a method executed by a computer system that improves the accuracy of sorting of objects in an image. The method extracts objects of a category in an image and reduces noise in the category. A second image is generated from the image with the category of objects in white color and the background in black color. A marked object is constructed in the second image by a pixel with at least two non-zero neighbor pixels. The number of the marked objects in the second image is counted.
Other example embodiments are discussed herein.
2018100159 05 Feb 2018
BRIEF DESCRIPTION OF THE DRAWINGS
Figure 1 shows a method that sorts objects in an image with an improved accuracy in accordance with an example embodiment.
Figure 2 shows an image in accordance with an example embodiment.
Figure 3 shows a preprocessed image in accordance with an example embodiment.
Figure 4 shows an image of extracted rocks in accordance with an example embodiment.
Figure 5 shows an image of categorized rocks in accordance with an example embodiment.
Figure 6A shows an image of extracted rocks in accordance with an example embodiment.
Figure 6B shows an image of extracted rocks with reduced noises in accordance with an example embodiment.
Figure 7 shows an image of rocks after clustering in accordance with an example embodiment.
Figure 8 shows an image of marked rocks in accordance with an example embodiment.
Figure 9 shows a computer system that executes the method with an example embodiment.
2018100159 05 Feb 2018
DETAILED DESCRIPTION
Example embodiments relate to methods and systems that extract characteristic information of the objects in an image and improve accuracy of identifying and sorting these objects.
Example embodiments provide examples of images as being linear scanning images. Example embodiments, however, are not limited to linear scanning images or a particular type of image. Various types of digital images can be recognized by a computer system or electronic device and can be executed as part of an example embodiment.
Linear scanning imaging has been applied in geological sample examination as well. Analyzing linear scanning images requires categorization and statistics of numerous irregular shaped objects, such as different kind of rocks in the image. The characteristics information including number, average size and size distribution of interested objects need to be extracted. Conventional techniques are time consuming and error prone due to the large amount of objects in each image, and the irregular shape and uneven color of objects.
Example embodiments solve one or more of the technical problems as set forth above by providing methods and systems that achieve automatic sorting of objects in an image with improved accuracy over conventional techniques. The objects are sorted into different categories and the information of objects in each category, including the number and average size of the objects, are extracted with a high accuracy. This information provides a more accurate, truer assessment of the geological sample, such as the material in the rock, the kind of rock, the composition of the rock, etc. This information helps scientists better understand the composition of geological samples.
In one example embodiment, firstly a category is defined in an image. Objects of the category are extracted and the noise in the category is reduced by a computer system or an electronic device executing an example embodiment.
A second image is generated from the image with the category of objects in
2018100159 05 Feb 2018 white color and the background in black color. A marked object is constructed in the second image by a pixel with at least two non-zero neighbor pixels. The number of the marked objects in the second image is counted and the number of pixels in each marked object is obtained. Finally, the statistic information of the objects in a category can be calculated from the pixel set of each marked object.
Figure 1 shows a method 100 that sorts objects in an image with an improved accuracy in accordance with an example embodiment.
Block 110 states defining a category in an image. By way of example, the category is defined according to the types of interested objects. In another example, the standard to classify the objects are provided by an image interpreter.
Block 120 states extracting objects of the category. By way of example, the objects of each category are exacted with HSV (Hue, Saturation, Value) model or RGB (Red, Green, Blue) model.
Block 130 states reducing noise in the category. In an example embodiment, the noise is reduced by applying a medium filter for individual RGB layer. In another example embodiment, the objects in a category are clustered by a clustering algorithm to reduce the errors.
Block 140 states generating a second image with the category of objects in white color and the background in black color. By this step, the objects of each category are shown in a separate binary image.
Block 150 states improving accuracy of sorting objects in the image, including constructing a marked object by a pixel with at least two non-zero neighbor, and counting a number of the marked objects in the binary image.
Block 160 states constructing a marked object by a pixel with at least two nonzero neighbor.
2018100159 05 Feb 2018
Each pixel has eight neighboring point. By way of example, if a non-zero pixel has at least two non-zero neighbors, a marked object is constructed with the connected non-zero pixels. The location of each marked object and the number of pixels therein are recorded at the same time.
Block 170 states counting a number of the marked objects in the binary image
Since the number and size of objects in each category is known, other 10 statistic information such as an average size can also be obtained from the pixel set of each marked object. As a result, the objects in the image are successfully sorted.
Figure 2 shows an image 200 in accordance with an example embodiment.
By way of example, the image 200 is a linear scanning image of a drilling sample in Chicxulub Crater. Analyzing the linear scanning image requires categorization and statistics of the irregular shaped rocks in this image.
For example, three categories of rocks are identified as interested objects and they are differentiated with their main color as white, black and green respectively. The characteristics information for each category, such as the number of interested objects, average size and size distribution of the objects are to be extracted.
Figure 3 shows a preprocessed image 300 in accordance with an example embodiment.
In an example embodiment, a pre-processing procedure is included to improve the quality of identifying the objects of different categories, since the linear scanning image may be taken with different lighting conditions, resulting in inconsistent brightness and contrast and that brings difficulty for sorting the objects in different images.
2018100159 05 Feb 2018
In one example, histogram equalization is used to alter the brightness and improve the image contrast. The image in Figure 3 shows the linear scanning image preprocessed by histogram equalization. The pre-processing procedure benefits the clustering of the objects in the images with sharp color difference among different types of objects.
Figure 4 shows an image 400 of extracted rocks in accordance with an example embodiment.
By way of example, exacting the objects of each category is achieved with HSV (Hue, Saturation, Value) model or RGB (Red, Green, Blue) model. For example, the categories of black and white stones are extracted with HSV model, and the green stone is extracted with RGB model.
In an example embodiment, HSV values of each pixel in the image is calculated with the HSV model. The pixel of the HSV value within a threshold of a certain category is classified as the corresponding category of rock. As shown in Figure 4, the white rocks in the image are extracted by HSV model.
Figure 5 shows an image of categorized rocks in accordance with an example embodiment. The categories of black, white and green stones are extracted and shown in different colors.
Figure 6A shows an image 600A of extracted rocks in accordance with an example embodiment. Figure 6B shows an image 600B of extracted rocks with reduced noises in accordance with an example embodiment.
Due to the uneven colored rock, the extracted objects may miss partial area or be classified into a wrong category. A filtering technique is used to remove noise from the image. In an example embodiment, a medium filter is applied on each individual RGB layer to mitigate the errors. All the filtered RGB layers are combined to form a new colored image. By this way, the unevenness within the object is smoothed.
2018100159 05 Feb 2018
As shown in Figure 6A, the black rocks are extracted with HSV model, which are shown as bright objects in a dark background. Figure 6B shows an image generated by applying medium filter on the image in Figure 6A. As can be seen in Figure 6B, the noises caused by non-uniform color of the rocks are reduced in the image.
Figure 7 shows an image 700 of rocks after clustering in accordance with an example embodiment.
In an example embodiment, a clustering algorithm is applied to the images with categorized black, white and green rocks to further reduce the errors of categorization. In each image, the recognized objects are shown in white color and the rests are in black. In one example, k-means algorithm is used as the clustering method. As shown in Figure 7, the green rocks are recognized after applying k-means clustering method and shown as white color in a black background.
Figure 8 shows an image 800 of marked rocks in accordance with an example embodiment.
In one example embodiment, the recognized objects of each category are shown as white color in a binary image. Each pixel in the binary image has eight neighboring points. A connected component is constructed if there is a non-zero pixel with at least two non-zero neighbors. The location of the pixels in the connected component are marked as well. The total number of the marked objects can be obtained by counting the number of connected components in the image.
In one example embodiment, a threshold is set to limit the size of the connected component to further remove the noises caused by different color of the same object. The connected component with an area smaller than the threshold is removed from the image and not counted as a marked object. Figure 8 shows the marked black rocks as white color in an image after applying a threshold.
2018100159 05 Feb 2018
Based on the number of marked objects and the number of pixels therewithin, characteristic information can be calculated. For example, an average size of the objects can be calculated by sqrt(A/(pi*N)), where sqrt is square root, A is the total area occupied by objects of a certain category, N is the number of the objects in the category.
In an example embodiment, the accuracy of the sorting method is 95%.
Figure 9 shows a computer system 900 that executes the method of sorting objects in an image with an improved accuracy in accordance with an example embodiment. The computer system 900 include one or more computer/server 910, a network 920 and a database 930.
The computer/server 910 includes a processor 911, a memory 912, a display 913 and an improved object sorter 914. The database 930 includes electronic storage or memory and can store data or other information to assist in executing example embodiments. The network 920 can include one or more of a wired network or wireless network, such as the internet.
The processor, memory and the improved object sorter in the computer/server 910 execute methods in accordance with example embodiments. The improved object sorter 914 can include software and/or hardware to execute example embodiments.
The processor unit includes a processor (such as a central processing unit, CPU, microprocessor, microcontrollers, field programmable gate array (FPGA), application-specific integrated circuit (ASIC), etc.) for controlling the overall operation of memory (such as random-access memory (RAM) for temporary data storage, read only memory (ROM) for permanent data storage, and firmware). The processing unit communicates with the improved object sorter and the memory to perform operations and tasks that implement one or more example embodiments discussed herein. The memory, for example,
2018100159 05 Feb 2018 stores applications, data, programs, algorithms (including software to implement or assist in implementing example embodiments) and other data.
In some example embodiments, the methods illustrated herein and data and instructions associated therewith are stored in respective storage devices, which are implemented as computer-readable and/or machine-readable storage media, physical or tangible media, and/or non-transitory storage media. These storage media include different forms of memory including semiconductor memory devices such as DRAM, or SRAM, Erasable and
Programmable Read-Only Memories (EPROMs), Electrically Erasable and Programmable Read-Only Memories (EEPROMs) and flash memories; magnetic disks such as fixed and removable disks; other magnetic media including tape; optical media such as Compact Disks (CDs) or Digital Versatile Disks (DVDs). Note that the instructions of the software discussed above can be provided on computer-readable or machine-readable storage medium, or alternatively, can be provided on multiple computer-readable or machine-readable storage media distributed in a large system having possibly plural nodes. Such computer-readable or machine-readable medium or media is (are) considered to be part of an article (or article of manufacture). An article or article of manufacture can refer to any manufactured single component or multiple components.
Blocks and/or methods discussed herein can be executed and/or made by a user, a user agent (including machine learning agents and intelligent user agents), a software application, an electronic device, a computer, firmware, hardware, a computer system, and/or an intelligent personal assistant. Furthermore, blocks and/or methods discussed herein can be executed automatically with or without instruction from a user.
The methods in accordance with example embodiments are provided as examples, and examples from one method should not be construed to limit examples from another method. Further, methods discussed within different figures can be added to or exchanged with methods in other figures. Further yet, specific numerical data values (such as specific quantities, numbers,
2018100159 05 Feb 2018 categories, etc.) or other specific information should be interpreted as illustrative for discussing example embodiments. Such specific information is not provided to limit example embodiments.
As used herein, “histogram equalization” is a method in image processing of contrast adjustment using the image's histogram.
As used herein, “HSV (Hue, Saturation, Value)” is a color model that describes colors (hue) in terms of their shade (saturation or amount of gray) and brightness (value)
As used herein, “RGB (Red, Green, Blue)” is an additive color model in which red, green and blue light are added together in various ways to reproduce a broad array of colors.
As used herein, “medium filter” is a nonlinear digital filtering technique used to remove noise from an image by replacing each entry with the median of its neighboring entries.
As used herein, “k-means” is a method for cluster analysis in data mining to partition n objects into k clusters in which each object belongs to the cluster with the nearest mean.
2018100159 05 Feb 2018
Claims (6)
- What is claimed is:1. A method executed by a computer system that improves the accuracy of sorting of objects in an image, the method comprising:defining a category;extracting, by the computer system, objects of the category; reducing, by the computer system, noise in the category; generating, by the computer system, a second image with the category of objects in white color and the background in black color; and improving the accuracy of sorting of objects in the image by:constructing, by the computer system, a marked object by a pixel with at least two non-zero neighbor pixels;counting, by the computer system, a number of the marked objects in the category.
- 2. The method of claim 1, further comprising:setting, by the computer system, a threshold; removing, by the computer system, the marked object with pixels less than the threshold.
- 3. The method of claim 1, further comprising:altering, by the computer system, brightness and/or contrast of the image.
- 4. The method of claim 1, further comprising:filtering, by the computer system, noise in the image.
- 5. The method of claim 1, further comprising:clustering, by the computer system, the objects in the image.1/92018100159 05 Feb 2018Fig.12/92018100159 05 Feb 2018Fig.23/92018100159 05 Feb 2018Fig.34/92018100159 05 Feb 2018Fig.45/92018100159 05 Feb 2018Fig.5
- 6/92018100159 05 Feb 2018600A600BFig.6BFig.6A2018100159 05 Feb 2018Fig.78/92018100159 05 Feb 2018Fig.89/92018100159 05 Feb 2018Figure 9
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
AU2018100159A AU2018100159A4 (en) | 2018-02-05 | 2018-02-05 | Method of sorting objects in an image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
AU2018100159A AU2018100159A4 (en) | 2018-02-05 | 2018-02-05 | Method of sorting objects in an image |
Publications (1)
Publication Number | Publication Date |
---|---|
AU2018100159A4 true AU2018100159A4 (en) | 2018-03-15 |
Family
ID=61568923
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
AU2018100159A Ceased AU2018100159A4 (en) | 2018-02-05 | 2018-02-05 | Method of sorting objects in an image |
Country Status (1)
Country | Link |
---|---|
AU (1) | AU2018100159A4 (en) |
-
2018
- 2018-02-05 AU AU2018100159A patent/AU2018100159A4/en not_active Ceased
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109154978B (en) | System and method for detecting plant diseases | |
US10438096B2 (en) | Identifying and excluding blurred areas of images of stained tissue to improve cancer scoring | |
JP4870363B2 (en) | High compression image data file generation device, method and program having a plurality of foreground planes | |
Agrawal et al. | Grape leaf disease detection and classification using multi-class support vector machine | |
US9740957B2 (en) | Learning pixel visual context from object characteristics to generate rich semantic images | |
Niu et al. | Image segmentation algorithm for disease detection of wheat leaves | |
CN112132843A (en) | Hematoxylin-eosin staining pathological image segmentation method based on unsupervised deep learning | |
CN112215790A (en) | KI67 index analysis method based on deep learning | |
WO2011084142A1 (en) | Image segregation system with method for handling textures | |
Masood et al. | Plants disease segmentation using image processing | |
KR101813223B1 (en) | Method and apparatus for detecting and classifying surface defect of image | |
CN103295013A (en) | Pared area based single-image shadow detection method | |
CN105844675A (en) | Color cluster analysis method of yarn-dyed fabric | |
CN115294377A (en) | System and method for identifying road cracks | |
Pandit et al. | Vision system for automatic counting of silkworm eggs | |
CN108171683B (en) | Cell counting method adopting software for automatic identification | |
AU2018100159A4 (en) | Method of sorting objects in an image | |
Salazar-Reque et al. | An image processing method to automatically identify Avocado leaf state | |
Ciobanu et al. | Image categorization based on computationally economic LAB colour features | |
CN115082741B (en) | Waste textile classification method based on image processing | |
Tarafdar et al. | Image segmentation using background subtraction on colored images | |
CN104581379A (en) | Video preview image selecting method and device | |
CN111160320A (en) | Image processing method | |
KR100488014B1 (en) | YCrCb color based human face location detection method | |
CN111563536B (en) | Bamboo strip color self-adaptive classification method based on machine learning |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
FGI | Letters patent sealed or granted (innovation patent) | ||
MK22 | Patent ceased section 143a(d), or expired - non payment of renewal fee or expiry |