Description:Complete Specification:
The following specification describes and ascertains the nature of this invention and the manner in which it is to be performed:

 
Field of the invention:
[0001] The present invention relates to a control unit and a method of reconstructing at least one unfocused image in a device

Background of the invention:
[0002] An innovative approach for obtained focused high-quality images from a digital microscopic system which has dynamic variations across samples has been discussed in the present invention. The device uses a stepper motor for movement of an image capturing unit and whenever there is a change in direction there is a certain amount of energy spent in compensating the reversal of polarity which leads to lesser movement than expected movement. This leads to precision loss in movement. Precise movement for microscopic applications is generally achieved by using piezo or hybrid stepper motors or motors with encoders using which the precise position can be determined. The back-clash compensation and re-capturing if required add significant amount of time in trying to correct the focus of image using hardware tuning.

[0003] According to state of the art US 9568724, a microscope including an objective having a focal plane in a sample space, and an autofocus device comprising a light modulator for generating a luminous modulation object that is intensity-modulated periodically along one direction, an autofocus illumination optical unit that images the modulation object such that its image arises in the sample space, an autofocus camera, an autofocus imaging optical unit that images the image of the modulation object in the sample space onto the autofocus camera, a control device, which receives signals of the autofocus camera and determines an intensity distribution of the image of the modulation object and generates a focus control signal therefrom.
Brief description of the accompanying drawings:
[0004] An embodiment of the disclosure is described with reference to the following accompanying drawings,
[0005] Fig. 1 illustrates a control unit in a device for reconstructing at least one unfocused image, according to an embodiment of the present invention; and
[0006] Fig. 2 illustrates a flowchart of a method of reconstructing of at least one unfocused image, according to the present invention.
Detailed description of the embodiments:
[0007] Fig. 1 illustrates a control unit for reconstructing of at least one unfocused image in a device according to one embodiment of the present invention. The device 10 comprises an image capturing unit 14 and an image processing unit 16. The device 10 further comprises a slot 18 for receiving a sample 20. The control unit 12 adjusts the image capturing unit 14 in a focal plane above a predefined distance from the sample 20. The control unit 12 captures a set of images of the sample 20 by the image capturing unit 14 and identifying the at least one unfocused image from the set of images and process the identified unfocused image in the image processing unit 16. The control unit 12 reconstructs the at least one focused image into a clear focused image using at least one intelligent module 22.

[0008] Further the construction of the device and the component of the device is explained in detail. According to one embodiment of the invention, the device 10 is a digital pathology device and the sample 20 is loaded with any one of the human fluids comprising blood, urine, saliva, spine fluid, semen and the like. Here for better understanding of the invention, the sample 20 is loaded with blood of a human being and the device is used to identify different types of cells in the sample 20. For instance, the red blood cells, the white blood cells and the platelets.

[0009] The image capturing unit 14 captures the images of the sample 20 in different planes comprising an x-axis plane, a y-axis plane and a z- axis plane. According to one embodiment of the invention, the image capturing unit 14 which is a camera captures the images in the z-axis plane. The axis plane referred here is also called as the focal plane. The image capturing unit 14 is made to stand at a predefined distance from the sample 20 while capturing the images. For instance, the predefined distance is ± 12.5μm from the focal plane. However, the predefined distance can be varied based on the requirement.

[0010] The images captured by the image capturing unit 14 is a combination of the unfocused images and focused images. The set of the images comprises at least one unfocused image that needs to be reconstructed. In the conventional method, the device 10 takes around “x” sec for capturing a clear focused image.

[0011] In order to analyze the white blood cell type, the device 10 takes a long time for capturing those clear focused images. For example, for analyzing the white blood cell type, the device 10 needs to capture around 500 clear focused images. The device 10 takes around 25sec for capturing one focused image. In order to complete the capture of 500 clear focused images, the device 10 will take around 2 and half hour period. In order to reduce the time taken for capturing those set of images of the sample 20, the present invention provides a solution in the form of reconstructing the unfocused images into focused images.

[0012] The intelligence module 22 present in the device 10 is trained with plurality of unfocused images and focused images during calibration. These images are referred as the pre-loaded images. The intelligence module 22 is trained in such a way that, when an unfocused image is given as an input, it reconstructs and provides a focused image by using the pre-loaded images and by pixel-to-pixel comparison technique. The intelligence module 22 is chosen from a group of modules comprising an artificial intelligence module, a deep learning module, a machine learning module, a generative adversarial module and the like. However, it is to be noted that, the intelligence module type need not to be restricted to the above-mentioned modules, but can be any other intelligence module that is known to a person skilled in the art.

[0013] The intelligence module 22 uses at least one neural network techniques along with the pixel-to-pixel comparison technique to reconstruct the unfocused image to a focused image. For instance, at least one unfocused image from the set of captured images is compared with all the pre-loaded images. By using the pixel-to-pixel comparison technique, the intelligence module 22 of the control unit 12 reconstructs the features of the images. Each pixel of the unfocused image is compared with each pixel of the images that are pre-loaded in the module 22. The intelligence module 22 reconstructs at least one feature in the identified unfocused image, wherein the at least one feature is a shape of a cell, color of a cell and the like. The pre-loaded images that are used to train the intelligence module 22 are stored in a memory 24 of the control unit 12.

[0014] Figure 2 illustrates a method of reconstructing at least one unfocused image in a device according to the present invention. The device 10 comprises an image capturing unit 14, an image processing unit 16 and a slot 18 for receiving a sample 20. The method involves the following steps. In step S1, the image capturing unit 14 in a focal plane is adjusted above a predefined distance from the sample 20. In step S2, a set of images of the sample 20 is captured by the image capturing unit 14. In step S3, the at least one unfocused image is identified from the set of images and the identified unfocused image is processed in said image processing unit 16. In step S4, the at least one focused image is reconstructed into a clear focused image using at least one intelligent module 22.

[0015] The above disclosed method is explained in detail. The digital pathology device 10 having the slot 18 receives the sample 20 that has the human fluid filled in. According to one embodiment of the invention, the human fluid is the blood. The blood sample is analyzed further to find any abnormality in the health of a patient. The device 10 captures the images of the blood present in the sample 20 for segregating the types of cells and for further analyzing the features of the segregated cells in the process of identifying the abnormality. The content of the sample is captured using the image capturing unit 14 and the captured images are processed further into the image processing unit 16. With the proposed method, the device 10 takes lesser time to complete a set of images that need to be captured, as the set combines the unfocused images and the focused images. All the unfocused images are identified by the control unit 12 and is processed further for reconstruction. In order to capture this set of images, the image capturing unit 14 is positioned in such a way that, it is held in a z-plane and above a predefined distance from the sample 20.

[0016] The control unit 12 upon identifying the unfocused images from the set of images, process each of the unfocused image via the intelligence module 22. The intelligence module 22 which has already multiple pre-loaded images (which are again a combination of the focused and unfocused images), compares each pixel of the unfocused image with the pixels of all the pre-loaded images for reconstruction of all the blur or unclear portions of the unfocused images.

[0017] For instance, the control unit 12 reconstructs the shape of the cells present in the unfocused part of the unfocused images by comparing each pixel of the unfocused image to each pixel of the each of the pre-loaded images. By that, the control unit 12 can identify the abnormality present in the person. The control unit 12 reconstructs multiple features of the cells present in the at least one unfocused image, wherein the feature of the cell is chosen from a group of features comprising a shape of the cell, color of the cell, type of the cell and the like. However, the features are not restricted to the above mentioned one’s, but can be of any other type that is known to a person skilled in the art.

[0018] With the above-mentioned method, the control unit 12 takes less time for providing the focused images for analyzing the content of the sample 20. As there are no additional components required, the method provides a cost-effective solution in saving the time of processing and makes the whole process more effective in a faster way. The disclosed method is compatible with entire visible spectrum and produces sharp full color RGB Images as output. The method is aimed at adding flexibility to the digital pathology device imaging hardware, since input required is any image in Z-plane within working range. The Z plane data is collected at each FOV (Field of View) in numerous slides covering almost all possible cell types and morphological representation. This data was used to train the constructed neural network in the intelligence module and optimized to converge where acceptable focused Image is obtained.

[0019] It should be understood that the embodiments explained in the description above are only illustrative and do not limit the scope of this invention. Many such embodiments and other modifications and changes in the embodiment explained in the description are envisaged. The scope of the invention is only limited by the scope of the claims.
, Claims:We claim:
1. A control unit (12) for reconstructing of at least one unfocused image in a device (10), said device (10) comprising an image capturing unit (14); an image processing unit (16); a slot (18) for receiving a sample (20), said control unit (12) adapted to:
- adjust said image capturing unit (14) in a focal plane above a predefined distance from said sample (20);
- capture a set of images of said sample by said image capturing unit (14);
- identify said at least one unfocused image from said set of images and process said identified unfocused image in said image processing unit (16);
- reconstruct said at least one focused image into a clear focused image using at least one intelligent module (22).

2. The control unit (12) as claimed in claim 1, wherein said image capturing unit (14) adapted to capture said set of images in a z-plane.

3. The control unit (12) as claimed in claim 1, wherein said set of images is a combination of unfocused images and focused images of said sample (20).

4. The control unit (12) as claimed in claim 1, wherein said identified at least one unfocused image is compared with each of a pre-loaded image that is stored in a memory (24) of said control unit (12).

5. The control unit (12) as claimed in claim 4, wherein said intelligence module (22) is trained to reconstruct said identified atleast one unfocused image from said set of pre-loaded images by a pixel-to-pixel comparison technique.

6. The control unit (12) as claimed in claim 1, wherein said device (10) is a digital pathology device and said intelligence module (22) is chosen from any one of modules comprising an artificial intelligence module, a deep learning module, a machine learning module, a generative adversarial module.

7. The control unit (12) as claimed in claim 1, wherein said intelligence module (22) reconstructs at least one feature in said identified unfocused image, wherein said at least one feature is a shape of a cell, color of a cell.

8. A method of reconstructing at least one unfocused image in a device (10), said device (10) comprising an image capturing unit (14); an image processing unit (16); a slot (18) for receiving a sample (20), said method comprising:
- adjusting said image capturing unit (14) in a focal plane above a predefined distance from said sample (20);
- capturing a set of images of said sample by said image capturing unit (14);
- identifying said at least one unfocused image from said set of images and process said identified unfocused image in said image processing unit (14);
- reconstructing said at least one focused image into a clear focused image using at least one intelligent module (22).