DESC:

FIELD OF THE INVENTION:
[0001] The present invention relates to images for Peripheral Blood Smear (PBS) slides, and more particularly relates to a system and method for capturing images from Peripheral Blood Smear (PBS) slides to locate good working area.

BACKGROUND OF THE INVENTION:
[0002] Blood Smear slides are used for detection of the pathogens in a blood specimen. The conventional way of finding a good working area on the slide involves manually scanning all the regions on the single slide having the blood smear.

[0003] To improve the detection of the pathogen in a peripheral blood smear slide an improved system having a single camera was proposed. The single camera captures the image of the peripheral blood smear slide and processes the image to diagnose the pathogen. However, the drawback of the single camera is that it takes very long as it scans the entire smear region instead of quick scanning the thin regions and then processing only the potential “good working areas” / monolayer regions. Further the image processing techniques as disclosed involvesuse of a hyperspectral camera. Further the output is restricted to the threshold of the wavelength.

[0004] The patent WO2009148363A2 discloses medicine, more specifically to the parasitic disease diagnosis which is based on blood test and can be used in clinical practice. The inventive method for diagnosing parasitic disease involves searching and identifying endoparasites and persistent microorganisms, i.e. infectious and parasitic disease causative agents, during the visualisation of the patient's native blood film via a microscope, which method differs in that the patient receives an antioxidant preparation, 20-30 minutes prior to the blood sampling, and in that the blood film is visualised via a microscope the magnification power of which is greater than 1000x. The invention makes it possible to increase the accuracy of the diagnosis of blood parasitic diseases and to reduce the diagnosis time.

[0005] The patent RU2098486C1 discloses a method for the diagnosis of bacteremia, including sampling the test blood, sowing and incubating it in a nutrient medium, followed by taking into account the grown colonies, characterized in that a leukoconcentrate isolated from the test blood is used for sowing.

[0006] The patent RU2123682C1 discloses method for detecting infectious and parasitic agents in analyzed drug involves producing standard preparation stained with video-specific fluorescent dyestuff, placing it on slide of fluorescent microscope stage, and illuminating the preparation with source suited to stimulate dyestuff fluorescence. Agents are searched while scanning the preparation in microscope field of vision with aid of stage followed by television sweep of microscopic images along two coordinates. Agents are identified by means of television camera which is coupled with microscope through video adapter. Computer used in the process is provided with program ensuring analysis of displayed images of agents and recording results in its storage. Facility implementing this method has fluorescent microscope with illuminating system in reflected light and system illuminating the preparation in incident light that scans microscope stage, two stepping motors, displacement system, image display system, and personal computer. Television camera of display system is mounted on microscope. Its output is connected to its interface control board through interface unit. Output of displacement system motor control unit is connected to stepping motors.

SUMMARY OF THE INVENTION:
[0001] In an aspect of the present disclosure, a system and a method for detection of a monolayer in a blood smear.

[0002] In one implementation, discloses real-time identification and detection of the monolayer region from a plurality of images. The implementation as disclosed comprises a dual camera technique which shortens the scanning region in the first step and performs detailed analysis only on the regions selected/shortlisted by the secondary camera.

[0003] In another implementation, a system for detecting monolayer having a microscope (300) is disclosed. The system further comprises a first/primary camera (302) mounted on the eyepiece of the microscope. Further a processor coupled to the first/primary camera (302). A secondary camera (304) may be mounted on a platform, wherein the platform may be mounted on the microscope (300) in accordance with the present disclosure. Further at least two light sources (308) are mounted on the platform. A plate (306) mounted on a stage (310), wherein the stage (310) is positioned at the distal end of the microscope from the eyepiece in accordance with the system.

[0004] In yet another implementation a method for detecting a monolayer in a blood smear is disclosed. The method may comprise capturing an image of the blood smear, using a primary and secondary camera. Further the method comprises detecting a monolayer in the blood smear using a monolayer detection model. Further at least one region is determined and may be called as the region of interest having higher probability of monolayer.

BRIEF DESCRIPTION OF DRAWINGS:
[0005] The detailed description is described with reference to the accompanying figures.
[0006] Figure 1, illustrates an exemplary embodiment in accordance with the present disclosure.
[0007] Figure 2 illustrates a system in accordance with the present disclosure.
[0008] Figure 3, illustrates a system in accordance with the present disclosure.
[0009] Figure 4, illustrates a flow chart in accordance with the present disclosure.
[0010] Figure 5, illustrates a flow chart for training the Monolayer Detection model in accordance with the present disclosure.

DETAILED DESCRIPTION OF THE INVENTION:
[0011] Reference throughout the specification to “various embodiments,” “some embodiments,” “one embodiment,” or “an embodiment” means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. Thus, appearances of the phrases “in various embodiments,” “in some embodiments,” “in one embodiment,” or “in an embodiment” in places throughout the specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures or characteristics may be combined in any suitable manner in one or more embodiments.

[0012] In accordance to an exemplary embodiment a system and a method is disclosed for detection of a monolayer in a periphery blood smear slide using a plurality of cameras and microscope.

[0013] The present disclosure discloses a primary camera and a secondary camera. The primary camera may be used to capture a monolayer from the detected “region of interest” from the peripheral blood smear (PBS) slides.

[0014] Referring to Figure 1 illustrates an embodiment in accordance with the present disclosure. According to the exemplary embodiment a system (100) comprises a microscope (106). The microscope (106) may be configured to magnify a specimen provided on a slide. The system (100) further comprises a primary camera (102). The primary camera (102) may be mounted on the microscope (106), and positioned near an eyepiece of the microscope (106). The primary camera (102) may be configured to capture the region of interest (ROI) under the observation of the slide. The ROI may be monolayer in case of peripheral blood smear slides. Further the region of interest (ROI) may be determined by a secondary camera (104).

[0015] Referring to Figure 2 illustrates a system in accordance with the present disclosure. The system as disclosed comprises a secondary camera, wherein the secondary camera may be configured to rapidly capture a plurality of good working area or region of interest (ROI) from the captured images of Peripheral Blood Smear (PBS) slides.

[0016] In accordance with the exemplary embodiment a whole slide image may be captured from the secondary camera to identify a probable monolayer region. Further a real-time deep learning algorithm filters the good working area determined from plurality of images captured of the whole slide, and discarding other images. Dynamic stepping based on the probability distribution from the deep learning algorithm is used for the movement of the motorized stage to reduce the scanning time of the defined region. The image processing algorithm may be executed once per slide, to determine and mark the probable monolayer region. Preprocessing of the image is done using image processing techniques to improve the detection of the monolayer region. Segmentation of the monolayer region is performed by the deep learning algorithm, yielding the physical location of the region. A mapping between this physical location and current stage position is produced to drive the motorized stage at the start of this region.

[0017] Given an image from probable monolayer region, a deep learning algorithm outputs a probability distribution over 6 types of images - blank, very sparse, sparse, good working area, clumped and very clumped.

[0018] The captured image is either stored for analysis or discarded, based on the probability distribution from the deep learning algorithm. This process of capturing and filtering images is repeated until at-least few 100 good working area images are captured.
[0019] Further in accordance with the exemplary embodiment the filtered images are uploaded to server. The filtered images are further mapped with a set of pre-defined images, wherein the based on the mapping, pathogens may be determined from the filtered image.

[0020] Now referring to Figure 3, illustrates a system in accordance with the present disclosure. The system as disclosed may comprise a microscope (300). Further the microscope (300) may comprise a first/primary camera (302) mounted on the eyepiece of the microscope to capture a region of interest i.e., a portion or region comprising monolayer of a blood smear on a slide (312). The primary camera (302) may be communicably coupled to a processor. The processor may further be configured to receive and send information, from and to, the primary camera (302) and a secondary camera (304). In accordance with the exemplary embodiment the processor may further be communicably connected with a storage module, and a detection module. A plurality of images captured by the primary camera (302) and the secondary camera (304) are stored in the storage module and further analysed for detecting monolayer by the detection module.

[0021] The secondary camera (304) may be mounted on a platform, wherein the platform may be mounted on the microscope (300). The secondary camera (304) may be configured to capture plurality of images to identify the region of interest. Further at least two light sources (308) are mounted on the platform, wherein the at least two light sources (308) may be focused on the slide (312) to reduce or avoid shadow effect. In accordance with the present disclosure the slide (312) may be mounted on a stage (310). Further a plate (306) may be mounted on the stage (310) on a side opposite to the slide (312). The plate (306) is positioned such that the slide (312) is positioned between the plate (306) and the primary camera (302). The plate (306) may be configured to enable better visibility of blood smear.

[0022] Referring to Figure 4, illustrates a flow chart in accordance with the present disclosure. In accordance with the flow chart, at step 402 an image of the blood smear is captured. Further at step 404, the processor is configured to communicate with the detection module to load a monolayer detection model. The monolayer model, further at step 406 determines at least one region called the region of interest having higher probability of having monolayer region. At step 408 an edited copy of the image is saved, wherein the edited copy has region of interest, i.e. having monolayer region.

[0023] Further at step 410, the edited image and the image not having the monolayer may be tagged as a function current region in the image and the probability of having monolayer. The images can be tagged as a blank region, a very sparse, a sparse, a monolayer, a clumped, and a very clumped region.

[0024] Further referring to Figure 5, illustrates a flow chart for training the Monolayer Detection model in accordance with the present disclosure. In accordance with the present disclosure assessing an image at high speed, a small neural network may be used. Further the small neural network may be trained using at least two big neural networks. This enables, the smaller neural network achieves high accuracy. Further the smaller neural network to assess the images.

[0025] The smaller neural network may be trained to be is both accurate and fast, however it may be slow to scan the whole slide. Further to reduce the slide scanning comprises preparing the slide such that the monolayer region is present approximately at the center (along the y-axis) of the slide. Further to avoid analysis of the surrounding region a dynamic step size may be used for the movement of the stage. Further the step size may be a function of the current region and output probabilities from the model. The probabilities for a certain region are highest after the start, and up to the center of that region. After this center of a region, the probabilities of that region start reducing because of the transition into the new region. Exploiting these properties, we can accurately skip over the initial blank, very sparse, and sparse regions using the dynamic step size formulation.
[0026] The embodiments, examples and alternatives of the preceding paragraphs or the description and drawings, including any of their various aspects or respective individual features, may be taken independently or in any combination. Features described in connection with one embodiment are applicable to all embodiments, unless such features are incompatible.

[0027] The foregoing description shall be interpreted as illustrative and not in any limiting sense. A person of ordinary skill in the art would understand that certain modifications could come within the scope of this disclosure. The embodiments, examples and alternatives of the preceding paragraphs or the description and drawings, including any of their various aspects or respective individual features, may be taken independently or in any combination. Features described in connection with one embodiment are applicable to all embodiments, unless such features are incompatible.
,CLAIMS:We Claim:

1. A system for detecting monolayer region having a microscope (300) comprises:
a. a first/primary camera (302) mounted on the eyepiece of the microscope;
b. a processor coupled to the first/primary camera (302);
c. a secondary camera (304) is mounted on a platform, wherein the platform is mounted on the microscope (300);
d. at least two light sources (308) are mounted on the platform; and
e. a plate (306) mounted on a stage (310), wherein the stage (310) is positioned at the distal end of the microscope from the eyepiece.

2. The system for detecting monolayer region having a microscope (300) as claimed in claim 1, wherein a portion or region comprising monolayer of a blood smear on a slide (312).

3. The system for detecting monolayer region having a microscope (300) as claimed in claim 1, wherein the processor is further communicably connected with a storage module.

4. The system for detecting monolayer region having a microscope (300) as claimed in claim 1, wherein the processor is further communicably connected with a detection module.

5. The system for detecting monolayer region having a microscope (300) as claimed in claim 1,wherein at least two light sources (308) may be focused on the slide (312) to reduce or avoid shadow effect.

6. The system for detecting monolayer region having a microscope (300) as claimed in claim 1, wherein the plate (306) is positioned such that the slide (312) is positioned between the plate (306) and the primary camera (302).

7. The system for detecting monolayer region having a microscope (300) as claimed in claim 1, wherein the plate (306) is configured to enable better visibility of blood smear.

8. A method of detection of monolayer region in a blood smear comprising:
a. capturing an image of the blood smear, using a primary and secondary camera;
b. detecting a monolayer in the blood smear using a monolayer detection model; and
c. determining at least one region called the region of interest having higher probability of monolayer.

9. The method of detection of monolayer region in a blood smear as claimed in claim 8, storing an edited copy of the image, wherein the edited copy has region of interest.

10. The method of detection of monolayer region in a blood smear as claimed in claim 8, wherein tagging the edited image and the image not having the monolayer as a blank region, a very sparse, a sparse, a monolayer, a clumped, and a very clumped region.

11. The method of detection of monolayer region in a blood smear as claimed in claim 8, wherein the monolayer detection model comprises training a small neural network using at least two big neural networks.

12. The method of detection of monolayer region in a blood smear as claimed in claim 11, wherein enabling the smaller neural network achieves high accuracy.

13. The method of detection of monolayer region in a blood smear as claimed in claim 11, wherein using a dynamic step size to move the stage for avoiding analysis of a region around the region of interest.