CN1553165A - Automatic multispectral cell smear and analyzing instrument and method used for cervical cell analysis - Google Patents
Automatic multispectral cell smear and analyzing instrument and method used for cervical cell analysis Download PDFInfo
- Publication number
- CN1553165A CN1553165A CNA2003101116100A CN200310111610A CN1553165A CN 1553165 A CN1553165 A CN 1553165A CN A2003101116100 A CNA2003101116100 A CN A2003101116100A CN 200310111610 A CN200310111610 A CN 200310111610A CN 1553165 A CN1553165 A CN 1553165A
- Authority
- CN
- China
- Prior art keywords
- image
- cell
- smear
- low power
- computing machine
- 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.)
- Granted
Links
Images
Abstract
An analyser comprises microscope, smear automatic placing device and control system, lightening system, computer, display, planar array CCD and data collection system, focusing device, automatic objective table, three-dimensional movable control system, electric-tuning light filter and its controller. The method includes a) program of collecting multioptical spectrum image and comprehensive processing program to carry on process from low power to high power for smear, and b) program of processing frow low power to high power to comprehensive and analysis for obtaining optical spectrum image.
Description
Technical field
The present invention relates to a kind of multispectral cell smear automatic analyzer and be used for the analytical approach of cervical cell, belong to the analysis and detecting instrument class, relate in particular to the medical inspection instrument that the cervical cell smear is analyzed automatically.
Technical background
The incidence of disease of cervical carcinoma occupies the first place of gynecologic malignant tumor, and the cervical carcinoma incidence of disease of China occupies the second place of the world.Pasteur's uterine neck is smeared the important content that the sheet inspection is the cervical carcinoma early diagnosis, as a kind of successful cervical carcinoma screening technique, it can provide and have or not canceration, a large amount of useful informations of the kind of inflammation and the order of severity, this method has made the cervical carcinoma mortality ratio of China obviously descend since using.But the uterine neck application of smearing sheet at present still rests on the artificial cognition stage: uterine neck is smeared sheet dye with Papanicolaou, pair cell carries out artificial one by one naked eyes inspection and examines under light microscopic, by Pasteur's Pyatyi classification the malignant tumour of female genital tract is diagnosed.This method workload is big, and every slice, thin piece on average has a hundreds of thousands cell, moves microscope with manual machinery entirely, visual inspection, very loaded down with trivial details and tired; Hospital expends great amount of manpower for this reason, and because doctor's the fatigue and the restriction of subjective experience etc., judges by accident and fail to pinpoint a disease in diagnosis still inevitable, has higher false negative rate and false positive rate.
There are following problems in external existing cell automated interpretation system: 1＞, because in the cell smear, the epithelial cell born of the same parents be difficult to avoid overlapping and cause nucleus and cytoplasmic cut apart inaccurate.2＞, the contradiction between recall rate and the false positive rate.Because adopted computer scanning to handle and new cytological classification method, the false negative rate of detection has reduced really, but false positive rate does not decline to a great extent.From 50,000,000 examples that the american computer inspection is examined, susceptibility is 90-97%, and the efficient of detection still paced up and down between 72-78%.3＞, for the cell that detects, system can not judge whether real unusual, still must manual detection confirm.The generation of these problems is all very few relevant with the image information that computing machine obtains.
Multi-optical spectrum imaging technology on the biomedicine can be used to strengthen the information of obtaining, and strengthens the recognition capability of computing machine pair cell.In principle, when Normocellular character changes, some materials in its nucleus, content as DNA (deoxyribonucleic acid) (DNA) also becomes thereupon, cause nuclear form, size and caryoplasm ratio to change, and the type that changes is according to the kind of cell and grade malignancy and different; Thereby, can only obtain the identifying information of cell 60%-70% at most to general cell image gray scale and morphologic analysis; Can not show a candle to multispectral image in the observation of cell form, by the imaging spectral analysis, obtain forming the information that changes in the cell, these cytochemistry component quantifying analytical informations make the original computer image analysis method that very application prospects arranged.From the document of existing application of spectral research tumour, with spectrum imaging system and the combination of cell microscopic analysis system, judge by spectrum whether cell inner structure and composition solve the present problem of system unusually, also be feasible.
Light spectrum image-forming comes from the remote sensing technology of the mid-80.Spaceborne (airborne) pull-broom type spectrum imaging system makes us can go out the most material in the face of land on the remote sensing images according to the characteristic spectrum Direct Recognition of atural object, therefore in geological prospecting, vegetation and crops generaI investigation and even military affairs purposes is widely arranged all.
Early begin cell smear computer diagnosis research abroad, and more existing utility system commercializations.PAPNET computer digital slide somascope as the U.S.: this device is a giant brain automatic scanning system, can imitate artificial intelligence highly difficult image is analyzed.Total system is made up of inspection center and intermediate cell chamber, can build two different places respectively in.In inspection center a computer microscan instrument and high quick kinescope are arranged, the intermediate cell chamber has a cover computer to separate picture system and microscope; The inspection of PAPNET is examined in two steps and is carried out, by computing machine each cell of smearing on the sheet is carried out preliminary identification screening in inspection center earlier, find out 128 the most suspicious cells from every slice, thin piece, record is taken with high quick kinescope in the visual field that will contain this cell, then in the intermediate cell chamber, by the cytologist data of record is reviewed, met shadow of doubt, carry out visual check at microscopically again; Cell is distinguished by the TBS classification by this system, exceeds much than traditional light microscopic screening accuracy rate.
Domestic research in this respect also has certain progress.Developed the computer micro-image disposal system cooperatively as University Of Tianjin in 1997 and Ji County, Tianjin healthcare hospital for women ﹠ children.This system focuses on the processing of optical signalling and the computer control aspect of microscope stage, the cervical cell image cut apart and identification aspect research less, thereby all there are some problems in the applicability of system and stability, practical application is less, and only hundreds of example cervical cell smear was examined in inspection.In Hubei Province, Sun Xiaorong is introduced the full-automatic cell image analyzer of Canadian BC Cancer Agency and has been set up Wuhan blue fourth early diagnosis of tumor inspection center, carry out the early stage work of checking of cancer that the cytology quantitative test is checked, the system that introduces adopts feulgen's stain, and mainly the dna content with cell serves as to detect foundation.
Do not see the report that multi-optical spectrum imaging technology is applied to the identification of cervical cell micro image analysis at present both at home and abroad as yet.
Summary of the invention
The invention provides a kind of automatic analyzer and multispectral cervical smear automatic analysis method of cell smear.This automatic analyzer method can increase the precision of cell detection and improve the accuracy rate that abnormal cell detects, and can satisfy the needs that clinical cytology is checked robotization.
The technical scheme of a kind of multispectral cell smear automatic analyzer provided by the invention is: it comprises microscope 1, automatic last slice device 3, last slice control system 4 automatically, illuminator 5, computing machine 10, display 12, area array CCD 15 and data acquisition system (DAS) 14; It also has: a, a zoom lens control device 2 are used for realizing suspicious or area-of-interest detects and whether high power belongs to the abnormal cell detection to suspicious cells to low power; B, automatic carrier 6 and three-dimensional mobile control system 7 and last slice device 3 automatically are used to realize the autoscan and the focusing of cell smear; C, electric tuning light filter 16 and an electric tuning control device of the light filter 8 and an area array CCD 15 that is configured on the microscope 1 are used to realize the quick collection of microspectrum image; Described electric tuning light filter 16 1 ends are installed on the microscope 1, the other end is connected with area array CCD 15, be imaged on the area array CCD 15 behind the stain smear cell picture process electric tuning light filter 16,10 in computing machine is regulated the electric signal that is added on the electric tuning light filter 16 by control control device of the light filter 8, change the centre frequency of its passband, area array CCD 15 detected picture signals convert digital picture to by image pick-up card and are stored in the computing machine 10, carry out Flame Image Process and analysis for computing machine 10.
The present invention is a kind of to be used for the technical scheme of the analytical approach of cervical cell with multispectral cell smear automatic analyzer, specifically, is the program of system; It comprises the steps: a, gathers multispectral image; B selects, the analytical multi-spectral image; C, measurement cell parameters; D, pair cell carry out pattern-recognition; E, comprehensive improvement analysis result; Described step is embodied in the program of system, comprise: a, a program of gathering multispectral image, it is the image that takes out different spectral bands after each regional area with cell smear amplifies through microscope 1 again by electric tuning light filter 16 respectively, obtains this zone after sending into computing machine through area array CCD 15, is the required spectrum picture of analysis of visual field; B, one by low power to high power at last to overall treatment, analyze the program of the spectrum picture that obtains, it is that multispectral image is carried out repeated screening, select interested image, accurately be partitioned into the karyon and the endochylema of cell, measure cell parameters, carry out pattern-recognition, last comprehensive whole analysis result; It is comprising the program of obtaining of a low power and high power image, and low power and high power are handled the program of monokaryon, multinuclear and overlap kernel cell, the program that low power cell image and high power cell image are cut apart.
One of maximum characteristics of effect of the present invention are the spectral characteristics of fully, rationally having utilized cell, make for cytoplasm and nuclear cut apart more accurate, speed is faster, broken away from the influence of dyeing time to analytical parameters and result simultaneously, the spectrum parameter of cell provides abundanter information for cell recognition; The analytic process of imitation doctor pair cell smear, system at first carries out pre-service under low power, remove impurity, bubble, suspicious zone is found in the influence of edge to analyzing, and analyzes under high power then, quickens analysis speed, is another one characteristics of the present invention.This technique effect also is: the automaticity height, and analysis result is accurate, and good reliability makes full use of and micro-ly multispectral cervical cell is carried out spectrum identification and classification, for the cervical disease diagnosis provides new spectroscopy foundation.
Description of drawings
Accompanying drawing 1 is the composition synoptic diagram of automatic analyzer system.Among the figure: 1-microscope, 2-zoom lens control device, 3-be last slice device automatically, 4-is last slice control system automatically, 5-illuminator, 6-automatic carrier, the three-dimensional mobile control system of 7-, 8-electric tuning control device of the light filter, 9-printer and other peripheral hardware, the 10-computing machine, 11-data backup restoration system, 12-display, the 13-picture monitor, the 14-data acquisition system (DAS), 15-area array CCD, 16-electric tuning device.
Accompanying drawing 2 is to utilize electric tuning to gather multispectral procedure chart in the system.
Accompanying drawing 3 be one by low power to high power at last to overall treatment, analyze the procedure chart of the spectrum picture that obtains, be the total handling procedure of system, begin by 100, move to required low power visual field 102, gather low power multispectral image 104, low power analyzes 106, termination condition 108, move to the subregion visual field 110 of location, low power visual field, gather high power multispectral image 112, high power analyzes 114, termination condition 116, overall treatment 118 is finished 120 and is formed.The system of summarizing that says is handled by low power, and high power is handled, and overall treatment is finished.
Accompanying drawing 4 is low power image processing programs in the system, by moving to low power visual field 122, visual field satisfaction evaluation 124, image segmentation 126, feature calculation 128, target classification 130, output information 132 compositions.
Accompanying drawing 5 is high power image processing programs in the system, and high power is handled by moving to satisfied high power visual field 140, image segmentation 142, and feature calculation 144, target classification 146, output information 148, termination condition 150 forwards next low power visual field 152 to and forms.
Accompanying drawing 6 is that system is divided into three classes with cell, and it was located the program synoptic diagram.
Accompanying drawing 7 is that image divided the program synoptic diagram in the system.
Accompanying drawing 8 is to adopt the multispectral design sketch of cutting apart in the system.
Embodiment
Accompanying drawing 1 is that a kind of multispectral cell smear automatic analyzer of the present invention system forms synoptic diagram.It has a microscope 1, automatic last slice device 3, last slice control system 4 automatically, illuminator 5, computing machine 10, display 12, area array CCD 15 and data acquisition system (DAS) 14; Also have a zoom lens control device 2, be used for realizing suspicious or area-of-interest detects and whether high power belongs to the abnormal cell detection to suspicious cells low power; An automatic carrier 6 and three-dimensional mobile control system 7 and last slice device 3 automatically are used to realize the autoscan and the focusing of cell smear; Electric tuning light filter 16 and an electric tuning control device of the light filter 8 and an area array CCD 15 that is configured on the microscope 1 is used to realize the quick collection of microspectrum image; Described electric tuning light filter 16 1 ends are installed on the microscope 1, the other end is connected with area array CCD 15, be imaged on the area array CCD 15 behind the stain smear cell picture process electric tuning light filter 16,10 in computing machine is regulated the electric signal that is added on the electric tuning light filter 16 by control control device of the light filter 8, change the centre frequency of its passband, area array CCD 15 detected picture signals convert digital picture to by image pick-up card and are stored in the computing machine 10, carry out Flame Image Process and analysis for computing machine 10; Computing machine 10 control is last slice control system 4 automatically, automatically the last slice automatic last slice device of control system control makes wave plate upload to automatic carrier 6, automatically last slice control system 4 detects and gives computing machine 10 signals after wave plate uploads to objective table, the three-dimensional mobile control system 7 of computing machine 10 control automatic carriers, the three-dimensional mobile control system 7 of automatic carrier makes wave plate be in the position that microscope 1 requires by control automatic carrier 6, and carries out automatic focus.Computing machine 10 control electric tuning control device of the light filter 8, data acquisition system (DAS) 14 is gathered multispectral image.Computing machine 10 makes microscope 1 automatically switch under low power and high power by control zoom lens control device 2.10 pairs of images that collect of computing machine are handled according to first low power, and back high power is handled, the method for last overall treatment, thus the sample on the wave plate is made suitable evaluation;
Automatic analyzer of the present invention has following additional technical feature on above-mentioned enforcement basis embodiment is:
Zoom lens control device 2 its zoom multiples are: the object lens slewing area of microscope 1 is a 4-20 low power doubly when analyzing beginning, and after the low power analysis was finished, the object lens slewing area of microscope 1 was a 20-100 high power doubly;
Automatically last slice device 3 is connected with computing machine 10 by last slice control system 4 automatically, by computing machine 10 controls automatically last slice device 3 loading cell smears to automatic carrier 6 and from automatic carrier 6, unload and analyzed the cell smear of finishing;
Automatic carrier 6 is connected with computing machine 10 by the three-dimensional mobile control system 7 of automatic carrier, the image calculation focus control automatic carrier of being gathered by area array CCD 15 by computing machine 10 6 moves up and down and carries out automatic focus, and control automatic carrier 6 moves horizontally realizes smear scanning;
The bandwidth of electric tuning light filter 16 is 2-30nm, preferred acousto-optic tunable filter.A kind of embodiment that multispectral cell smear automatic analyzer is used for the analytical approach of cervical cell of the present invention: as Fig. 1～shown in Figure 8, it is made up of following step: a, gather multispectral image; B selects, the analytical multi-spectral image; C, measurement cell parameters; D, pair cell carry out pattern-recognition; E, comprehensive improvement analysis result; Described step is embodied in the program of system: a, a program of gathering multispectral image, it is the image that takes out different spectral bands after each regional area with cell smear amplifies through microscope 1 again by electric tuning light filter 16 respectively, obtains this zone after sending into computing machine through area array CCD 15, is the required spectrum picture of analysis of visual field; B, one by low power to high power at last to overall treatment, analyze the program of the spectrum picture that obtains, it is that multispectral image is carried out repeated screening, select interested image, accurately be partitioned into the karyon and the endochylema of cell, measure cell parameters, carry out pattern-recognition, last comprehensive whole analysis result; The program of obtaining that a low power and high power image are wherein arranged, low power and high power are handled the program of monokaryon, multinuclear and overlap kernel cell, the program that low power cell image and high power cell image are cut apart;
Measure the parameter of cell, comprise morphological parameters, parametric texture and the spectrum parameter of nuclear and slurry, carrying out pattern-recognition is according to the relation between cytoplasm and nucleus; Divide three kinds of situations to measure the parameter of cell: 1, monocyte, a nucleus is only contained in a cytoplasm the inside, and cytoplasm and nucleus all have good classifying quality; 2, apocyte, a cytoplasm the inside is contained two nucleus at least, but mutual adhesion between the nucleus, nucleus has good separating effect; 3, overlap kernel cell; A cytoplasm the inside is contained two inter-adhesive overlapping nucleus at least;
Accompanying drawing 2 is the programs of gathering multispectral image:
Owing to spectrum picture is combined by a series of monochrome image, therefore, need under each our wavelength selected, gather piece image under each visual field.Open light source and electric tuning control device of the light filter, behind light stability half an hour, cervical smear is placed on objective table approximately, light source is adjusted to a suitable position, all be put into identical position later at every turn, make illumination intensity identical for identical enlargement ratio.Regulate microscope to kohler's illumination, even with the brightness that guarantees the visual field.
The program of gathering multispectral image is: a, under each visual field, at first the transmission peak wavelength of electric tuning light filter 16 is navigated to the middle wavelength of the wavelength that will gather, under the control of computing machine 10, utilize image processing algorithm to carry out automatic focus, make the visual field the most clear; B, switch to one by computing machine 10 control electric tuning light filters 16 and do not gather under the wave band of selecting; Gather piece image the integral time that c, usefulness calculate; D, the image that collects is carried out gray correction; E, when selected wave band end scan, and then switch to next selected wave band, gather piece image again, carry out gray correction again, so circulation finishes up to the whole collections of all selected wave bands; F, carry out again that colour is synthetic at last, Treatment Analysis and data storage.
The program of gray correction is: focus on smear, make cell the best clear.A, mobile smear, defocus under an empty visual field slightly to white space, by computing machine 10 control electric tuning light filter 16 and area array CCDs 15, with gathering piece image at each selected wave band an identical integral time; B, calculate each wave band next suitable integral time according to these images again, a long integral time can be given for the wave band that gray-scale value is lower, give integral time of a weak point for the high wave band of gray-scale value, to carry out balanced compensated to whole wavelength band; Again gather piece image, spectrum picture at the bottom of the back of the body that is enhanced the integral time that c, basis calculate again under each selected wave band; D, later image for each wave band all use gather the integral time identical with spectrum picture at the bottom of the back of the body that strengthens, and gray-scale value is promoted, and guarantee that all there is consistent gray-scale value in the back bottom area territory at different wave bands, is convenient to Flame Image Process and analysis.Because the correction to the back of the body end is that unit handles with the pixel, so not only can effectively remove the imaging system influence different to different-waveband sensitivity, also can compensate lifting by a pair of blue wave band, effectively eliminate some slight visual field unevenness and area array CCD 15 each pixel luminous sensitivity difference.
Accompanying drawing 3 be one by low power to high power at last to overall treatment, analyze the procedure chart of the spectrum picture that obtains, be the total handling procedure of system, the total processing of system is handled by low power, the high power processing, overall treatment is finished.The computing machine 10 of a, system guarantees that at first the image that obtains from area array CCD 15 can be truly or near the multispectral information that truly reflects cervical smear.Because traditional doctor follows at the check cervical smear: at first inspect below low-power microscope, find suspicious zone, forward to then under the high-power microscope each cell of suspicious zone is carried out careful examination, find suspicious cell, thereby the judgement of closing the symbol diagnostics is made in section.According to same principle, the low power processing is also followed in the processing of smear to native system and high power is handled principle of combining; B, open low power image for each of arriving of detecting of area array CCD 15, the algorithm software of system computer 10 control system computing machines is handled it, judge whether this image contains suspicious target, the suspicious target of handling in low power of process kind comprises suspicious unusual monocyte, suspicious apocyte; C, in case detect and have suspicious target in this image, system computer will be controlled microscope and forward high power to, and the subregion that has suspicious object is carried out meticulous screening; D, in case detect the low power image and do not contain interested suspicious object or do not have interested visual field, system computer will carry out the low power image acquisition process of next to smear; According on the method chatted, computing machine under low power to smear on all targets carry out scan process, satisfy up to certain condition, for example: detected cell number reaches certain upper limit; Detect the suspicious cells number and reach certain number; What system detected reaches certain upper limit T.T.; In case system handles finds suspicious target in the low power visual field, system computer will be preserved the position of suspicious object and microscope forwarded under the high power suspicious object will be handled accurately; Because the visual field that microscope is obtained under low power is handled and between the visual field that is obtained under the high power, there are differences, briefly the zone scanned, low power visual field is bigger than the zone that high power scanned, so system is the approximate subregion that is divided into some in the low power visual field of containing suspicious object; Under high power, system is according to certain priority ranking, for example: what of suspicious object subregion contain; What of target subregion contain; Or the like, successively scan these subregions; If subregion does not contain interested target, system-computed will forward the subregion of next order to; E, computing machine carry out scan process to all sub-visual fields under high power, up to satisfying certain condition, for example: satisfy system handles and finish condition, satisfy all subregions and scanned condition or the like, f, in the overall treatment stage 118, system will fully utilize low power analysis result 106, and high power analysis result 114 is assessed smear.The result of assessment comprises: normal or need rescan.
Accompanying drawing 4 is programs of low power Flame Image Process.Traditionally for the processing of cervical cell or be based on gray level image, or be colored RGB triple channel image, but practice shows that the information that traditional images provides very little for the cervical cell image of complexity.On method, the method for general using threshold value utilizes three values that the cervical cell image is carried out dividing processing.High threshold provides the background information of image, the slurry information of the cell that middle threshold value provides, and low threshold value provides the nuclear information of image.Yet adopting the method is to be based upon on the ideal situation basis, and it is very low that promptly slurry, background, the nuclear in the cervical cell image is subjected to actual conditions interference such as impurity, noise.In fact be subjected to the influence of actual operating conditions, on the ideal situation chatted be difficult to occur, for most cervical cell image, contrast difference is nowhere near for the method that adopts threshold process between background, slurry, the nuclear.The accuracy and the consistance reason of the work that will determine the later stage based on the accuracy and the consistance of image segmentation, native system utilizes the feature of multispectral image that the uterine neck image is cut apart.System will cut apart each width of cloth image in the low power processing procedure, be partitioned into interesting target.Interesting target comprises: monocyte, apocyte.In the practice because normally whether the doctor judges a smear, comprehensive often micro details and macroscopical overall situation information, the doctor should judge accurately to individual cells, also will be considered to the situation of blocking cell.In to the smear evaluation process,, can not ignore the booster action that apocyte and overlap kernel cell provide though monocyte plays most important effect to assessment.So native system is for becoming the blocking cell also to give enough attention.System at first assesses 124 to the low power visual field.Handle the visual field for low power and following certain situation may occur: (a) this visual field is a blank, that is to say that this visual field do not contain any target; (b) bubble phenomenon is contained in this visual field, because bubble belongs to impurity and interference fully for the uterine neck Flame Image Process, it is owing to improper pre-service causes to cervical smear that bubble occurs, and belongs to pretreated problem, so must get rid of bubble phenomenon in the visual field; (c) because the even visual field, slivering zone of being caused of uneven illumination.For the situation that may occur in above three, system computer must exclude earlier when low power is handled.Because it is to determine suspicious zone interested and some overall information are provided that low power is handled significant feature, so system has only utilized the fraction of 36 wave bands altogether when low power is handled.When handling, low power saved the time of a large amount of samplings like this, the processing time.In the image segmentation stage 126, system will make full use of multispectral feature to be cut apart accurately to cervical cell, detailed describing see accompanying drawing 7 with and explanation.For cytoplasm that separates and nuclear, feature calculation 128 will be calculated the series of features value.These eigenwerts had both comprised the morphological feature (obtaining from overlayer) of image, also comprised object content feature (overlayer institute overlay area target obtains), and a series of images global characteristics also will calculate simultaneously.Because image that low-power microscope obtained is especially little, thus it is accurately measured impossible, at this time for the texture of target, the contextual measurement particular importance that just seems.In 130 stages of target classification, to train for the various cells that may occur in the image according to the priori system, the cell of training has: normal monocyte, unusual monocyte, suspicious apocyte, normal apocyte, various impurity.The result system that makes of training can be good at telling unusual monocyte, suspicious apocyte.According to the result of training, system design goes out a cover layering arbiter, and every layer all comprises the normal monocyte of eliminating, impurity, normal apocyte.The layering arbiter, it is every that one deck is all accurate than last layer down, and the feature that every certainly one deck is down measured is also all many than last layer.Native system has designed three layers arbiter.Final system can be discerned those and contain potential unusual monocyte, potential unusual many cells.At information output stage 132, the low power analysis will be carried out information evaluation to each low power visual field of handling.These evaluations comprise: the positional information in suspicious object zone, visual field macroscopic information, various cell information.According to the suspicious object area information, system will control microscope and forward the high power processing to.Describe in detail and see accompanying drawing 5 and explanation thereof.
Accompanying drawing 5 is programs of high power Flame Image Process.High power is handled similar low power and is handled.Because when low power was handled, the image that is obtained was smaller, so its processing accuracy is not high comparatively speaking; And high power is when handling, and institute's image that obtains is relatively big, and all types of target is also relatively big, so its relative accuracy will height.At first system handles according to low power and obtains the suspicious object zone position information, and the computer control microscope forwards under the high power, to the subregion automatic focus of target area, obtains multispectral image.In the image segmentation stage 142, system will make full use of multispectral information to be cut apart cervical cell, detailed process see accompanying drawing 7 with and explanation.In the feature calculation stage, system carries out various pattern measurements with independent cell to three types, and these measurements had both comprised the various morphological features of target, also comprised the content characteristic of target.In the target classification stage 146, system classifies to three class cells according to feature extraction 142: normal monocyte, unusual monocyte, normal apocyte, unusual apocyte, normal overlap kernel cell, unusual overlap kernel cell.Termination condition 150 is set, whether finishes, and get back to next visual field of low power suitable the time and handle to judge high power scanning.Termination condition is for example: handle all high power visual fields; System handles is finished.Output information 148 will be preserved (with file or other modes) various useful informations in high power visual field, and these information comprise: unusual monokaryon information, unusual apocyte information, unusual overlap kernel cell information and visual field information.
Accompanying drawing 6 is that low power processing and high power are handled the handling procedure synoptic diagram to three kinds of cells: in cervical smear manufacturing process; usually the cell overlap phenomenon can appear; though use certain tabletting technology (for example ThinPrep ) to alleviate overlapping degree to a certain extent, it is impossible thinking thoroughly to eliminate film-making process kind cell overlap.Based on the relation of slurry in the cell and nuclear, native system is divided into type in three with cell, monocyte, apocyte, overlap kernel cell.The reason that cell is divided into three types is: 1＞, for monocyte,, segmentation effect is very accurately arranged because cytoplasm and nuclear all are released states, and do not exist the adhesion situation between the same kind, the most accurate cell information is provided.2＞, for apocyte, owing to exist overlapping between the cytoplasm, the nucleus that does not have adhesion in every cytoplasm, it is impossible that each cytoplasm is cut apart accurately, and cytoplasmic information can not reflect the real situation of cell fully, and native system is used as the cytoplasm of these adhesions as an integral body and is handled, but nucleus all has good separating effect, and real information more accurately can be provided.3＞, for the overlap kernel cell, because all there are adhesion in slurry and nuclear, so still all can not reflect its truth fully for the dividing processing of nuclear for slurry, at this moment an integral body is used as the cytoplasm of adhesion and the nuclear that wherein comprises as by system.According to the top characteristics of describing three class cells, obviously this three classes cell has different segmentation precisions and measuring accuracy, and different measurement features that produces therefrom and the effect different to the contribution of system, so thereby system with the cell fully reasonable use cell information that is divided three classes.For each class cell, system will independently carry out parameter measurement, pattern-recognition, cell evaluation.For monocyte, the cell evaluation comprises: normal monocyte number, unusual monocyte number.For apocyte, the cell evaluation comprises: can detected all multi-nucleus cell numbers, can detected suspicious unusual multi-nucleus cell number.For the overlap kernel cell, the cell evaluation comprises: can detected overlap kernel cell number, can detect suspicious unusual overlap kernel cell number.In comprehensive evaluation, system is with comprehensive three kinds of evaluation results, and utilization discriminatory analysis (discriminatory analysis) pair cell carries out comprehensive evaluation.Thereby smear is provided evaluation.
Accompanying drawing 7 is procedure charts of cutting apart for low power cell image and high power cell image in the system, is to utilize multi-band image that cervical cell is cut apart, and comprising: a, select some interested band images from multi-band image; B, the logarithm ratio that utilizes the band image of selecting or ratio produce new image; C, cut apart comprise pair cell slurry cut apart nuclear cutting apart; D, cut apart and produce nuclear overlayer, the cytoplasm overlayer; E, the uninterested target of eliminating or zone; Dividing method comprises the following formula of use:
In the formula: I----is the light intensity of target; I
N-----newly-generated image; ρ
NThe absorptivity of----target.Describe its principle and dividing method below in detail.According to Lambert law: I=I
0e
-ρ dWherein I is the light intensity (target) of target, I
0Be incident intensity (background) that ρ is the absorptivity of target, d is the thickness of target.In conjunction with multispectral information, derive:
Formula 6-1
By formula 6-1 explanation, newly-generated image I
NOnly and ρ
NAbsorptivity is relevant, and with incident intensity, the thickness of target is irrelevant.Interference (the I that obviously can get rid of incident light by the newly-generated image of formula 6-8
0), the interference of target overlapping (d) phenomenon.Newly-generated image I
NBy ρ
1And ρ
2Decision, and ρ and material composition, filter wavelength is relevant, for fixed target (material composition is fixed), I
NOnly relevant with λ.It is the present invention's one big characteristic that application of formula 6-1 is cut apart the cervical cell image.By formula 6-1, some band images are at first selected according to training and mathematical statistics method by system, band image to these selections carries out the logarithm variation, image after logarithm changes carries out the new gray level image of combination results, new gray level image has strengthened the contrast of background and target greatly, got rid of the influence of illumination, the influence of cell overlap.System utilizes the method for threshold value to cut apart in new image basis, because new image has strengthened the contrast (the effect synoptic diagram is seen accompanying drawing 8a, 8b) of background and target greatly, so system can utilize a fine low target and background that separates of threshold value.Native system design cut apart the method that is not according to traditional, only image segmentation is become the 0-1 image, but produces a series of overlayers, utilize the unique expression target of overlayer.System will produce the nucleus overlayer, monocyte overlayer, apocyte overlayer, overlap kernel cell overlayer.Producing tectal benefit is: (a) overlayer can be shown the morphological feature of target; (b) overlayer can be shown the cell content characteristic.And simple traditional 0-1 image only can reflect is the morphological feature of cell.Utilize the background and the target image of Threshold Segmentation will produce the cytoplasm overlayer, this overlayer is with the unique identification cytoplasm.Because the cervical cell slurry tends to occur the unsound situation of slurry, starches into the cobweb form.Native system has utilized the method for a series of mathematical morphologies, and for example: expansion, burn into opening operation, closed operation are smoothly sealed processing to slurry, handles the back slurry and will produce fine and close overlayer.Same method, some wave bands are selected by data training and mathematical statistics by system, the band image of these selections is carried out logarithm change, notice the wave band that selects this moment and wave band that slurry is handled also inconsistent.Adopt the similar method can nuclear and other targets is separated, similarly method be for the nucleus overlayer, and native system also adopts a series of mathematical morphologies to handle, and handles the back nucleus and will produce the nucleus overlayer of densification.Final system is with the cytoplasm overlayer, and the nucleus overlayer stacks up and finishes tentatively cutting apart of target.The overlayer of tentatively cutting apart only can the cytoplasmic information of unique identification cell nucleus, system will consider the situation that each cytoplasm contains nuclear this moment, if cytoplasm only contains a nucleus, system thinks that this cell is a monocyte so, system will redistribute overlayer, the information of overlayer unique identification monocyte endochylema and karyon this moment.If cytoplasm contains a plurality of nuclears, this cell apocyte will be thought by system so, same system will redistribute new overlayer, and this overlayer is noticed apocyte overlayer and monocyte overlayer color and inequality with the information of unique identification apocyte endochylema and karyon.
Claims (10)
1, a kind of multispectral cell smear automatic analyzer, it comprises microscope (1), automatic last slice device (3), last slice control system (4) automatically, illuminator (5), computing machine (10), display (12), area array CCD (15) and data acquisition system (DAS) (14); It is characterized in that a, a zoom lens control device (2) being arranged, be used for realizing suspicious or area-of-interest detects and whether high power belongs to the abnormal cell detection to suspicious cells low power; B, an automatic carrier (6) and three-dimensional mobile control system (7) and automatic last slice device (3) are used to realize the autoscan and the focusing of cell smear; C, electric tuning light filter (16) and an electric tuning control device of the light filter (8) and an area array CCD (15) that is configured on the microscope (1) are used to realize the quick collection of microspectrum image; Described electric tuning light filter (16) one ends are installed on the microscope (1), the other end is connected with area array CCD (15), be imaged on the area array CCD (15) behind the stain smear cell picture process electric tuning light filter (16), computing machine (10) is then regulated the electric signal that is added on the electric tuning light filter (16) by control control device of the light filter (8), change the centre frequency of its passband, the detected picture signal of area array CCD (15) converts digital picture to by image pick-up card and is stored in the computing machine (10), carries out Flame Image Process and analysis for computing machine (10).
2, by the described automatic analyzer of claim 1, it is characterized in that: its zoom multiple of zoom lens control device (2) is: the object lens slewing area of microscope (1) is a 4-20 low power doubly when analyzing beginning, after the low power analysis was finished, the object lens slewing area of microscope (1) was a 20-100 high power doubly.
3, by the described automatic analyzer of claim 1, it is characterized in that: automatic last slice device (3), be connected with computing machine (10) by automatic last slice control system (4), load cell smear by the automatic last slice device (3) of computing machine (10) control and upward and from automatic carrier (6) unloading analyzed the cell smear of finishing to automatic carrier (6).
4, by the described analyser of claim 1, it is characterized in that, automatic carrier (6) is connected with computing machine (10) by the three-dimensional mobile control system of automatic carrier (7), the image calculation focus control automatic carrier of being gathered by area array CCD (15) by computing machine (10) (6) moves up and down and carries out automatic focus, and control automatic carrier (6) moves horizontally realizes smear scanning.
5, by the described analyser of claim 1, it is characterized in that the bandwidth of electric tuning light filter (16) is 2-30nm, preferred acousto-optic tunable filter.
6, a kind of the described multispectral cell smear automatic analyzer of claim 1 is used for the analytical approach of cervical cell, its spy is that it comprises the steps: a, gathers multispectral image; B selects, the analytical multi-spectral image; C, measurement cell parameters; D, pair cell carry out pattern-recognition; E, comprehensive improvement analysis result; Described step is embodied in the program of system, comprise: a, a program of gathering multispectral image, it is the image that takes out different spectral bands after each regional area with cell smear amplifies through microscope (1) again by electric tuning light filter (16) respectively, passes through and obtains this zone after area array CCD (15) is sent into computing machine, is the required spectrum picture of analysis of visual field; B, one by low power to high power at last to overall treatment, analyze the program of the spectrum picture that obtains, it is that multispectral image is carried out repeated screening, select interested image, accurately be partitioned into the karyon and the endochylema of cell, measure cell parameters, carry out pattern-recognition, last comprehensive whole analysis result; Comprising the program of obtaining of a low power and high power image, low power and high power are handled the program of monokaryon, multinuclear and overlap kernel cell, the program that low power cell image and high power cell image are cut apart.
7, by the described method of claim 6, it is characterized in that, gathering the multispectral image program is: a, under each visual field, at first the transmission peak wavelength of electric tuning light filter (16) is navigated to the middle wavelength of the wavelength that will gather, under the control of computing machine (10), utilize image processing algorithm to carry out automatic focus, make the visual field the most clear; B, switch to one by computing machine (10) control electric tuning light filter (16) and do not gather under the wave band of selecting; Gather piece image the integral time that c, usefulness calculate; D, the image that collects is carried out gray correction; E, when selected wave band end scan, and then switch to next selected wave band, gather piece image again, carry out gray correction again, so circulation finishes up to the whole collections of all selected wave bands; F, carry out again that colour is synthetic at last, Treatment Analysis and data storage.
8, by the described method of claim 7, it is characterized in that, the program of described gray correction comprises: a, mobile smear are to white space, under an empty visual field, defocus slightly, by computing machine (10) control electric tuning light filter (16) and area array CCD (15), with gathering piece image at each selected wave band an identical integral time; B, calculate each wave band next suitable integral time according to these images again, a long integral time can be given for the wave band that gray-scale value is lower, give integral time of a weak point for the high wave band of gray-scale value, to carry out balanced compensated to whole wavelength band; Again gather piece image, spectrum picture at the bottom of the back of the body that is enhanced the integral time that c, basis calculate again under each selected wave band; D, later image for each wave band all use gather the integral time identical with spectrum picture at the bottom of the back of the body that strengthens, and gray-scale value is promoted, and guarantee that all there is consistent gray-scale value in the back bottom area territory at different wave bands, is convenient to Flame Image Process and analysis.
9, by the described method of claim 6, it is characterized in that, the pair cell smear carries out being handled to high power by low power, carries out the overall treatment program at last and comprises: a, computing machine (10) guarantee that at first the image that obtains from area array CCD (15) can be truly or near the multispectral information that truly reflects cervical smear; B, each low power image that detection obtains for area array CCD (15) are handled it by computing machine (10), judge whether this image contains suspicious target, comprise suspicious unusual monocyte, suspicious apocyte; C, in case detect and have suspicious target in this image, system computer (10) will be controlled microscope (1) and forward high power to, and the subregion that has suspicious object is carried out meticulous screening; D, in case detect the low power image and do not contain interested suspicious object or do not have interested visual field, system computer (10) will carry out the low power image acquisition process of next to smear; E, computing machine (10) carry out scan process to all sub-visual fields under high power, up to satisfying certain condition, can be to satisfy system handles to finish condition, or satisfying all subregions, to have scanned bar assorted; F, in the overall treatment stage, system will fully utilize the low power analysis result and the high power analysis result is assessed smear, the result of assessment comprises: normal or need rescan.
10, by the described method of claim 6, it is characterized in that, carrying out the dividing processing program for low power cell image and high power cell image in the system is to utilize multi-band image that cervical cell is cut apart, and comprising: a, select some interested band images from multi-band image; B, the logarithm ratio that utilizes the band image select or ratio produces new image: c, cut apart comprise the pair cell slurry cut apart nuclear cutting apart; D, cut apart and produce nuclear overlayer, the cytoplasm overlayer; E, the uninterested target of eliminating or zone; Dividing method comprises the use formula:
In the formula: I---is the light intensity of target; I
V-----newly-generated image; ρ
NThe absorptivity of----target.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200310111610 CN1236302C (en) | 2003-12-19 | 2003-12-19 | Automatic multispectral cell smear and analyzing instrument and method used for cervical cell analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200310111610 CN1236302C (en) | 2003-12-19 | 2003-12-19 | Automatic multispectral cell smear and analyzing instrument and method used for cervical cell analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1553165A true CN1553165A (en) | 2004-12-08 |
| CN1236302C CN1236302C (en) | 2006-01-11 |
Family
ID=34336220
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200310111610 Expired - Fee Related CN1236302C (en) | 2003-12-19 | 2003-12-19 | Automatic multispectral cell smear and analyzing instrument and method used for cervical cell analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1236302C (en) |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101339185B (en) * | 2008-06-23 | 2011-03-02 | 武汉呵尔医疗科技发展有限公司 | Automatic microscopic imager for detecting cast-off cells and detection method |
| CN102682305A (en) * | 2012-04-25 | 2012-09-19 | 深圳市迈科龙医疗设备有限公司 | Automatic screening system and automatic screening method using thin-prep cytology test |
| CN102788756A (en) * | 2012-07-13 | 2012-11-21 | 上海凯度机电科技有限公司 | Multi-modal biological microscope analyzer |
| CN103489188A (en) * | 2013-09-23 | 2014-01-01 | 华南理工大学 | Positioning method for abnormal cell nucleus in cervix uteri LCT image |
| CN103926687A (en) * | 2013-01-16 | 2014-07-16 | 索尼公司 | Image Acquisition Apparatus, Image Acquisition Method, And Information Processing Program |
| CN103975364A (en) * | 2011-12-05 | 2014-08-06 | 皇家飞利浦有限公司 | Selection of images for optical examination of the cervix |
| CN104459172A (en) * | 2014-11-28 | 2015-03-25 | 武汉兰丁医学高科技有限公司 | Automatic bone marrow sample processing device, and automated analyzing and radiograph reading method thereof |
| CN101331500B (en) * | 2005-10-14 | 2015-04-29 | 尤尼森斯繁殖技术公司 | Determination of a change in a cell population |
| CN104614332A (en) * | 2015-02-02 | 2015-05-13 | 武汉呵尔医疗科技发展有限公司 | Quantitative measurement method of cell DNA under redyeing environment |
| CN105247346A (en) * | 2013-05-29 | 2016-01-13 | 佳能株式会社 | Spectral microscopy device |
| CN105738365A (en) * | 2016-01-27 | 2016-07-06 | 马心峰 | Automatic dyeing and analyzing device for pathological diagnosis |
| CN106124420A (en) * | 2016-07-07 | 2016-11-16 | 威凯检测技术有限公司 | A kind of spectral detection instrument data analysis and information synthesis method |
| CN106595860A (en) * | 2016-11-27 | 2017-04-26 | 苏州国科美润达医疗技术有限公司 | Multispectral imaging system |
| CN107103604A (en) * | 2017-05-23 | 2017-08-29 | 重庆天之助生物科技有限公司 | A kind of particulate colourity auto-clustering analysis system |
| CN108982500A (en) * | 2018-07-03 | 2018-12-11 | 怀光智能科技（武汉）有限公司 | A kind of cervical liquid-based cells intelligence auxiliary diagosis method and system |
| CN110231259A (en) * | 2019-05-28 | 2019-09-13 | 怀光智能科技（武汉）有限公司 | A kind of cervical cell slide numerical dialing system |
| CN110411926A (en) * | 2019-07-31 | 2019-11-05 | 长沙协大生物科技有限公司 | A method of Non-Gaussian Distribution concentration of cell suspension is estimated by small sampled data |
| CN110530781A (en) * | 2019-08-30 | 2019-12-03 | 武汉呵尔医疗科技发展有限公司 | A kind of three band digital Image Acquisition stripping system of single frames |
| CN113139485A (en) * | 2021-04-29 | 2021-07-20 | 新乡医学院 | Bone marrow cell classification and identification method, device and system based on deep learning |
| CN113420653A (en) * | 2021-06-22 | 2021-09-21 | 深圳跃美生物医学科技有限公司 | Method, device and readable medium for comprehensive detection and analysis of bacterial cell morphology |
-
2003
- 2003-12-19 CN CN 200310111610 patent/CN1236302C/en not_active Expired - Fee Related
Cited By (27)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101331500B (en) * | 2005-10-14 | 2015-04-29 | 尤尼森斯繁殖技术公司 | Determination of a change in a cell population |
| CN101339185B (en) * | 2008-06-23 | 2011-03-02 | 武汉呵尔医疗科技发展有限公司 | Automatic microscopic imager for detecting cast-off cells and detection method |
| CN103975364A (en) * | 2011-12-05 | 2014-08-06 | 皇家飞利浦有限公司 | Selection of images for optical examination of the cervix |
| CN103975364B (en) * | 2011-12-05 | 2017-04-12 | 皇家飞利浦有限公司 | Selection of images for optical examination of the cervix |
| CN102682305A (en) * | 2012-04-25 | 2012-09-19 | 深圳市迈科龙医疗设备有限公司 | Automatic screening system and automatic screening method using thin-prep cytology test |
| CN102788756A (en) * | 2012-07-13 | 2012-11-21 | 上海凯度机电科技有限公司 | Multi-modal biological microscope analyzer |
| CN103926687A (en) * | 2013-01-16 | 2014-07-16 | 索尼公司 | Image Acquisition Apparatus, Image Acquisition Method, And Information Processing Program |
| CN105247346A (en) * | 2013-05-29 | 2016-01-13 | 佳能株式会社 | Spectral microscopy device |
| CN103489188A (en) * | 2013-09-23 | 2014-01-01 | 华南理工大学 | Positioning method for abnormal cell nucleus in cervix uteri LCT image |
| CN104459172A (en) * | 2014-11-28 | 2015-03-25 | 武汉兰丁医学高科技有限公司 | Automatic bone marrow sample processing device, and automated analyzing and radiograph reading method thereof |
| CN104614332A (en) * | 2015-02-02 | 2015-05-13 | 武汉呵尔医疗科技发展有限公司 | Quantitative measurement method of cell DNA under redyeing environment |
| CN104614332B (en) * | 2015-02-02 | 2017-05-10 | 武汉呵尔医疗科技发展有限公司 | Quantitative measurement method of cell DNA under redyeing environment |
| CN105738365B (en) * | 2016-01-27 | 2018-11-06 | 马靖华 | Pathological diagnosis automatic staining analytical equipment |
| CN105738365A (en) * | 2016-01-27 | 2016-07-06 | 马心峰 | Automatic dyeing and analyzing device for pathological diagnosis |
| CN106124420A (en) * | 2016-07-07 | 2016-11-16 | 威凯检测技术有限公司 | A kind of spectral detection instrument data analysis and information synthesis method |
| CN106595860B (en) * | 2016-11-27 | 2018-12-14 | 苏州国科美润达医疗技术有限公司 | Multi-optical spectrum imaging system |
| CN106595860A (en) * | 2016-11-27 | 2017-04-26 | 苏州国科美润达医疗技术有限公司 | Multispectral imaging system |
| CN107103604A (en) * | 2017-05-23 | 2017-08-29 | 重庆天之助生物科技有限公司 | A kind of particulate colourity auto-clustering analysis system |
| CN108982500A (en) * | 2018-07-03 | 2018-12-11 | 怀光智能科技（武汉）有限公司 | A kind of cervical liquid-based cells intelligence auxiliary diagosis method and system |
| CN108982500B (en) * | 2018-07-03 | 2020-07-14 | 怀光智能科技（武汉）有限公司 | Intelligent auxiliary cervical fluid-based cytology reading method and system |
| CN110231259A (en) * | 2019-05-28 | 2019-09-13 | 怀光智能科技（武汉）有限公司 | A kind of cervical cell slide numerical dialing system |
| CN110231259B (en) * | 2019-05-28 | 2022-02-18 | 怀光智能科技（武汉）有限公司 | Digital diagnosis system for cervical cell slide |
| CN110411926A (en) * | 2019-07-31 | 2019-11-05 | 长沙协大生物科技有限公司 | A method of Non-Gaussian Distribution concentration of cell suspension is estimated by small sampled data |
| CN110411926B (en) * | 2019-07-31 | 2022-03-08 | 长沙协大生物科技有限公司 | Method for estimating concentration of non-normally distributed cell suspension through small sampling data |
| CN110530781A (en) * | 2019-08-30 | 2019-12-03 | 武汉呵尔医疗科技发展有限公司 | A kind of three band digital Image Acquisition stripping system of single frames |
| CN113139485A (en) * | 2021-04-29 | 2021-07-20 | 新乡医学院 | Bone marrow cell classification and identification method, device and system based on deep learning |
| CN113420653A (en) * | 2021-06-22 | 2021-09-21 | 深圳跃美生物医学科技有限公司 | Method, device and readable medium for comprehensive detection and analysis of bacterial cell morphology |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1236302C (en) | 2006-01-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1236302C (en) | Automatic multispectral cell smear and analyzing instrument and method used for cervical cell analysis | |
| CN1296699C (en) | Microscopic multispectral marrow and its peripheral blood cell auto-analyzing instrument and method | |
| CN101339185B (en) | Automatic microscopic imager for detecting cast-off cells and detection method | |
| US6181414B1 (en) | Infrared spectroscopy for medical imaging | |
| US4175860A (en) | Dual resolution method and apparatus for use in automated classification of pap smear and other samples | |
| US6718053B1 (en) | Method and apparatus for automated image analysis of biological specimens | |
| CA1161271A (en) | Method and apparatus for measuring mean cell volume of red blood cells | |
| US7783098B2 (en) | Method and apparatus for automated image analysis of biological specimens | |
| US5848177A (en) | Method and system for detection of biological materials using fractal dimensions | |
| US20090245612A1 (en) | Automated image analysis | |
| EP0317139A2 (en) | Methods and apparatus for cell analysis | |
| CN112580748B (en) | Method for counting classified cells of stain image | |
| CN111062346A (en) | Automatic leukocyte positioning detection and classification recognition system and method | |
| JP2020169995A (en) | Biosensor platform and method for simultaneous, multiplexed, ultra-sensitive and high throughput optical detection of biomarkers | |
| US11519856B2 (en) | Method for optically detecting biomarkers | |
| WO2017079212A1 (en) | Cell detection, capture, analysis, aggregation, and output methods and apparatus | |
| DE102009033927B4 (en) | Apparatus and method for locating a valid region of a blood smear | |
| CN111398138A (en) | Optical detection system and method of dry type blood cell analysis device | |
| CN1522362A (en) | Method for analyzing a biological sample | |
| Ahmad et al. | Colorectal cancer image classification using image pre-processing and multilayer Perceptron | |
| CA2279945A1 (en) | Infrared spectroscopy for medical imaging | |
| CN108387553B (en) | Block reconstruction and classification counting method for leucocyte and platelet coexistence hologram | |
| Schneider et al. | Computer-aided cytological cancer diagnosis: cell type classification as a step towards fully automatic cancer diagnostics on cytopathological specimens of serous effusions | |
| CN115541578B (en) | High-flux super-resolution cervical cell pathological section rapid scanning analysis system | |
| Prasad et al. | A high throughput screening algorithm for leukemia cells |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C19 | Lapse of patent right due to non-payment of the annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |