Method and system to obtain images of target cells in cytopathology

Abstract

Example methods and systems to capture images of target cells distributed in a cytology specimen have been disclosed. One example method includes obtaining a first z-stack image and a second z-stack image, generating a first focus map, capturing images of the target cells according to the first focus map, and generating a second focus map. A first distance along a z-axis of the cytology specimen between a first set of target cells associated with the first focus map and a second set of target cells associated with the second focus map is different from a second distance along the z-axis of the cytology specimen between a third set of target cells associated with the first focus map and a fourth set of target cells associated with the second focus map.

Description

METHOD AND SYSTEM TO OBTAIN IMAGES OF TARGET CELLS IN CYTOPATHOLOGYCROSS-REFERENCE TO RELATED APPLICATION

[0001] This application claims the benefit of U.S. Provisional Application No. 63 / 652,051 , filed May 26, 2024, which is incorporated by reference in its entirety.BACKGROUND OF THE INVENTIONField of the Invention

[0002] Embodiments of the present invention generally relate to methods and systems to obtain images of target cells in cytopathology.Description of the Related Art

[0003] Unless otherwise indicated herein, the approaches described in this section are not prior art to the claims in this application and are not admitted to be prior art by inclusion in this section.

[0004] Cytopathology is a branch of pathology that studies and diagnoses diseases on a cellular level and generally involves obtaining cytology images on the cellular level. Obtaining cytology images of a cytology specimen involves digitizing images of the glass slide on which the cytology specimen is distributed. The image digitization may generally include scanning the glass slide with a whole-slide imaging scanner for the cytology specimen to generate a whole slide image (WSI). The WSI can be viewed on a display (e.g., computer monitor) instead of a microscope.

[0005] Capturing target cells on a glass slide efficiently and precisely via image digitization can be challenging. One conventional approach in digital imaging is to create one or more focus maps based on focal positions across multiple regions of the glass slide and attempt to capture images of target cells according to the focus maps. However, in the conventional approach, different focus maps are often set to maintain an arbitrary distance of separation among them without considering how target cells are distributed in a suspension of a cytology specimen, where cells can be distributed in three dimensions and at varying depths. Therefore, relying on conventional focus maps to capture target cells distributed on the glass slide isinefficient and inaccurate, resulting in poor-quality captured images of the target cells.BRIEF DESCRIPTION OF THE DRAWINGS

[0006] Fig. 1 is an example figure showing a cytology specimen distributed in a three-dimensional space on a glass slide, arranged in accordance with some embodiments of the present disclosure;Fig. 2 illustrates how an example scanner obtains images associated with a cytology specimen, arranged in accordance with some embodiments of the present disclosure;Fig. 3 illustrates an exploded view of a cytology specimen including various parts, arranged in accordance with some embodiments of the present disclosure;Fig. 4A illustrates a conventional exploded view of cytology specimen;Fig. 4B illustrates another conventional exploded view of cytology specimen;Fig. 5A illustrates an exploded view of cytology specimen, arranged in accordance with some embodiments of the present disclosure;Fig. 5B illustrates another conventional exploded view of cytology specimen, arranged in accordance with some embodiments of the present disclosure;Fig. 6 illustrates an example scanner configured to generate one or more focus maps across a cytology specimen and obtain images associated with target cells in the cytology specimen based on the focus maps, arranged in accordance with some embodiments of the present disclosure;Fig. 7 is a flow diagram illustrating an example process to obtain images associated with target cells distributed in a cytology specimen, arranged in accordance with some embodiments of the present disclosure; andFig. 8 is an example device configured to perform various embodiments of the present disclosure, arranged in accordance with some embodiments of the present disclosureDETAILED DESCRIPTION

[0007] In the following detailed description, reference is made to the accompanying drawings, which form a part hereof. In the drawings, similar symbols typically identify similar components, unless context dictates otherwise. The illustrative embodiments described in the detailed description, drawings, and claims are not meant to be limiting. Other embodiments may be utilized, and other changes may be made, without departing from the spirit or scope of the subject matter presented here. It will be readily understood that the aspects of the present disclosure, as generally described herein, and illustrated in the Figures, can be arranged, substituted, combined, and designed in a wide variety of different configurations, all of which are explicitly contemplated herein. In the description, a cytology specimen is suspicious when target cells distributed in a space of the cytology specimen include cells at risk of a disease.

[0008] In the disclosure, a cytology specimen includes multiple two-dimensional regions within the Cartesian x-y plane, and each of these two-dimensional regions is referred to as a “region.” A “z-stack image” of a region refers to a series of images for the region, captured at different focal positions of the region along the z-axis (i.e., depth of field) of the cytology specimen. A “focus map” refers to a three-dimensional representation of optimal focal positions across multiple regions of a cytology specimen.

[0009] Fig. 1 is an example figure showing cytology specimen 1 10 distributed in a three-dimensional space on glass slide 120, arranged in accordance with some embodiments of the present disclosure. Cytology specimen 110 may include multiple cells and impurities. For example, cytology specimen 1 10 may include dust or mark 131 , target cells 141 (e.g., the Epithelial cells at malignant risk), non-target cells 151 , 153, 155, 157 and 159 (e.g., red blood cells, normal Epithelial cells and other cells). Dust or mark 131 , target cells 141 and non-target cells 151 , 153, 155,157 and 159 are distributed in a three-dimensional space (i.e., at different depths of field within the volume of cytology specimen 110) on glass slide 120.

[0010] Fig. 2 illustrates how an example scanner 220 obtains images associated with cytology specimen 210, arranged in accordance with some embodiments of the present disclosure. In conjunction with Fig. 1 , in Fig. 2, cytology specimen 210 may correspond to cytology specimen 110. In Fig. 2, scanner 220 is configured to obtain images associated with one or more layers of cytology specimen 210. In some embodiments, each layer may correspond to a depth of field associated an object lens included in scanner 220.

[0011] In some embodiments, an object lens included in scanner 220 has a first field of view and a first depth of field. The first field of view may correspond to any of regions A, B, C, D, E, F, G, H, I, and J of cytology specimen 210.

[0012] In some embodiments, in the first field of view, a distance between a sharp and focused object nearest to scanner 220 and a sharp and focused object furthest to scanner 220 is referred to as the first depth of field. Some example distances may be d1 , d2, d3, and d4 illustrated in Fig. 2. Therefore, d1 may correspond to the first layer of cytology specimen 210; d2 may correspond to the second layer of cytology specimen 210; d3 may correspond to the third layer of cytology specimen 210; and d4 may correspond to the fourth layer of cytology specimen 210.

[0013] In conjunction with Fig. 1 , The first layer may be a layer furthest away from a glass slide (e.g., glass slide 120) and the fourth layer may be a layer adjacent to the glass slide (e.g., glass slide 120). Noting that although Fig. 2 illustrates 4 layers, cytology specimen 210 may include more or less layers according to the first depth of field of the object lens included in scanner 220. In these embodiments, a region and a distance may define a “part” of cytology specimen 210. For example, J3 isa part of cytology specimen 210 corresponding to region J and d3. Therefore, in one embodiment illustrated in Fig. 2, cytology specimen 210 may be defined by 40 parts, i.e., A1 , A2, A3, A4, B1 , B2, B3, B4, C1 , C2, C3, C4, D1 , D2, D3, D4, E1 , E2, E3, E4, F1 , F2, F3, F4, G1 , G2, G3, G4, H1 , H2, H3, H4, 11 , I2, I3, I4, J1 , J2, J3, and J4.

[0014] In some embodiments, Fig. 3 illustrates an exploded view of cytology specimen 300 including various parts, arranged in accordance with some embodiments of the present disclosure. In some embodiments, cytology specimen 300 may correspond to cytology specimen 210. In some embodiments, for illustration, parts A1 , B1 , C1 , D1 , E1 , F1 , G1 , H1 , 11 , and J1 correspond to the first layer of cytology specimen 300 and are at the same depth of field as illustrated in Fig. 3. In some embodiments, for illustration, parts A1 , A2, A3, and A4 correspond to different layers associated with the same region A of cytology specimen 300 along a Z-axis as illustrated in Fig. 3.

[0015] Referring back to Fig. 2, in some embodiments, cytology specimen 210 may include, but not limited to, dust 211 , mark 212, first type of cells 213 (e.g., the Epithelial cell at malignant risk), second type of cells 214 (e.g., red blood cell), third types of cells 215 (e.g., normal Epithelial cell) and fourth types of cells 216 (e.g., other cells). Dust 211 , first type of cells 213, second type of cells 214, third types of cells 215, and fourth types of cells 216 may be distributed in cytology specimen 210. Mark 212 may be manually marked by a user on the top surface of cytology specimen 210. Dust 211 and cells 213, 214, 215 and 216 may be distributed in different layers of cytology specimen 210. In some embodiments, cells 213 are target cells, and cells 214, 215, and 216 are non-target cells.

[0016] In some embodiments, prior to obtaining images associated with cytology specimen 210 with scanner 220, scanner 220 is configured to generate one or more focus maps across multiple regions of cytology specimen 210. Each focus map is used to guide scanner 220 to dynamically adjust the focus of scanner 220 as scanner 220 moves across different regions of cytology specimen 210, ensuring that each region of cytology specimen 210 can be imaged with the focal information specified in the focus map.

[0017] To generate a focus map, scanner 220 is configured to randomly select various regions across cytology specimen 210 and determine the optimal focal position for each selected region based on some imaging characteristics of the scanned region. Alternatively, scanner 220 is configured to select various regions across cytology specimen 210 based on information obtained in an earlier scan ofscanner 220 (e.g., a lower magnification scan of scanner 220) and determine the optimal focal position for each selected region based on some imaging characteristics of the scanned region. For illustration, scanner 220 may be configured to randomly select regions B, D, F, H, I, and J from all regions and obtain images associated with these selected regions. For example, scanner 220 is configured to obtain images of parts H1 , H2, H3, and / or H4 in the randomly selected region H. Here, region H corresponds to a first field of view of an object lens included in scanner 220, and d1 , d2, d3, or d4 corresponds to a first depth of field of the object lens. Similarly, scanner 220 may also obtain images of parts B1 , B2, B3, B4, D1 , D2, D3, D4, F1 , F2, F3, F4, 11 , I2, I3, I4, J1 , J2, J3, and J4 of cytology specimen 210.

[0018] Conventionally, scanner 220 is configured to determine the optimal focal position of a region based on contrasts, edges, and color values of images of parts associated with the region. Continuing with the example with region H above, after obtaining the images of parts H1 , H2, H3, and H4, scanner 220 is configured to compare the images of parts H1 , H2, H3, and H4 and determine which image has the highest contrast, the sharpest edge, or the brightest color value. In conjunction with Fig. 2, assuming the image of part H2 has the highest contrast among the images of parts associated with region H, scanner 220 is configured to determine the depth of field for part H2 to be the optimal focal position for region H. Similarly, for illustration, scanner 220 is configured to determine the depths of field for parts B2, D2, F2, I2, and J2 to be the optimal focal positions for regions B, D, F, I, and J, respectively. Fig. 4A illustrates a conventional exploded view of cytology specimen 400. In Fig. 4A, the depths of field for parts B2, D2, H2, F2, I2, and J2 are determined to correspond to the optimal focal positions of a first focus map in regions B, D, H, F, I, and J, respectively.

[0019] Conventionally, for regions not scanned by scanner 220 (i.e., non-scanned regions), scanner 220 is configured to estimate the optimal focal positions for these non-scanned regions (e.g., regions A, C, E, and G) based on the optical focal positions of the scanned regions (e.g., regions B, D, F, H, I, and J). The estimation may be based on some technical feasible approaches, such as linear or bilinearinterpolation, triangulation, or more advanced computational methods to generate a smooth, continuous primary focus map across the cytology specimen. Fig. 4A also illustrates a conventional exploded view of cytology specimen 400 in which the depths of field for parts A2, C2, E2, and G2 being estimated to be the optimal focal positions of the first focus map in regions A, C, E, and G, respectively. Given the first focus map is associated with the highest contrast, the sharpest edge, or the brightest color value, the first focus map may be also called a “primary focus map.” After the primary focus map is generated, scanner 220 is configured to focus all regions of the cytology specimen according to the primary focus map to obtain images of all regions of the cytology specimen.

[0020] Fig. 4B illustrates another conventional exploded view of cytology specimen 400. After generating the primary focus map, scanner 220 may be configured to generate one or more additional secondary focus maps. These secondary additional focus maps are generated by offsetting the optimal focal positions of the primary focus map by a fixed distance, allowing scanner 220 to capture images from different layers of cytology specimen 400. In other words, conventionally, such a secondary focus map is generated in parallel with the primary focus map, meaning that a distance between any corresponding focal position in the primary and secondary focus maps is kept constant throughout the entire cytology specimen. This approach aims to capture more comprehensive image information from various layers of the cytology specimen. The distance may be set by a user based on his / her experience. For illustration, in Fig. 4B, the depths of field for parts A4, B4, C4, D4, E4, F4, G4, H4, 14, and J4 may be determined to be the optimal focal positions of the secondary focus map in all regions A, B, C, D, E, F, G, H, I, and J of cytology specimen 400, respectively. After the secondary focus map is generated, scanner 220 is configured to focus all regions of cytology specimen 400 according to the secondary focus map to obtain additional images of all regions of cytology specimen 400.

[0021] However, conventional approaches have deficiencies. First, the primary focus map is generated based on the highest contrast, the sharpest edge, and / or the brightest color value. However, the highest contrast, the sharpest edge, and / or the brightest color value do not necessarily correspond to the presence of target cells.For example, in conjunction with Fig. 2 and Fig. 4A, target cells are mainly distributed in the third layer of cytology specimen 210. However, the primary focus map is generated at the second layer of cytology specimen 210. Accordingly, images of target cells distributed in cytology specimen 210 are less likely to be obtained based on the primary focus map. Second, the secondary focus maps are generated by offsetting the optimal focal positions of the primary focus map by a user-set fixed distance. However, the user-set fixed distance also fails to consider image characteristics and presence of target cells. For example, in conjunction with Fig. 2 and Fig. 4B, the secondary focus map is generated to correspond to the fourth layer of cytology specimen 210. However, there is no target cell distributed in the fourth layer of cytology specimen 210.

[0022] Fig. 5A illustrates an exploded view of cytology specimen 500, arranged in accordance with some embodiments of the present disclosure. In some embodiments, cytology specimen 500 may correspond to cytology specimen 300. In conjunction with Fig. 2, in some embodiments, scanner 220 is configured to generate a primary focus map across cytology specimen 500. More specifically, scanner 220 is configured to randomly select various regions across cytology specimen 210 and determine the optimal focal position for each scanned region based on some imaging characteristics of the region. More specifically, scanner 220 is configured to obtain z-stack images of several regions randomly selected from all regions A, B, C, D, E, F, G, H, I and J. For example, scanner 220 may be configured to randomly select regions B, D, F, H, I, and J and obtain z-stack images associated with these selected regions. More specifically, for illustration, scanner 220 is configured to obtain a first z-stack image associated with region B (i.e., images associated with parts B1 , B2, B3, and / or B4). Similarly, scanner 220 may also obtain a second, third, fourth, fifth, and sixth z-stack images associated with regions D, F, H, I, and J, respectively (i.e., images associated with parts D1 , D2, D3, D4, F1 , F2, F3, F4, H1 , H2, H3, H4, 11 , I2, I3, I4, J1 , J2, J3, and / or J4).

[0023] In some embodiments, the z-stack images are processed by scanner 220 with an artificial intelligence engine to identify whether the z-stack images include images of target cells 213. The artificial intelligence engine may include machine learningcapabilities. The artificial intelligence engine may be trained based on sample images of known cells corresponding to target cells 213 having various contrasts, lightness, shapes and other image characteristics. In some embodiments, scanner 220 is configured to compare images included in the same z-stack image and identify the image having the highest number of target cells in the same z-stack image. For example, in region F, scanner 220 is configured to compare images included in the third z-stack image (e.g., images of parts F1 , F2, F3, and F4) and identify the image of part F3 having the highest number of target cells in the third z- stack image as illustrated in Fig. 2. Similarly, referring back to Fig. 5A, scanner 220 is configured to identify images of parts B2, D2, H3, I3, and J3 having the highest number of target cells in the first, second, fourth, fifth, and sixth z-stack images, respectively. Therefore, scanner 220 is configured to determine the depths of field for parts B2, D2, F3, H3, I3, and J3 to be the optimal focal positions for regions B, D, F, H, I, and J, respectively.

[0024] In some other embodiments, instead of randomly selecting a region of cytology specimen 500, the artificial intelligence engine may identify where target cells 213 are in a region (e.g., region F) of cytology specimen 500. For example, the artificial intelligence engine may infer one or more bounding boxes for target cells 213. For each bounding box that meets a confidence score threshold, the artificial intelligence engine may obtain z-stack images within the bounding boxes (e.g., F1 , F2, F3, F4). For illustration, scanner 220 is configured to compare the images of parts F1 , F2, F3, and F4, determine the image of part F3 has the highest contrast, the sharpest edge, the brightest color value, or the sharpest edge among images of parts F1 , F2, F3, and F4, and determine the depth of field for part F3 to be the optical focal position for region B.

[0025] In some embodiments, for non-scanned regions not scanned by scanner 220, scanner 220 is configured to estimate the optimal focal positions for the non-scanned regions (e.g., regions A, C, E, and G) based on the optimal focal positions of the scanned regions (e.g., regions B, D, F, H, I, and J). The estimation may be based on some technical feasible approaches, such as linear or bilinear interpolation, triangulation, or more advanced computational methods to generate a smooth,continuous primary focus map across the cytology specimen. Fig. 5A also illustrates an exploded view of cytology specimen 500 in which the depths of field for parts A2, C2, E2, and G3 are estimated to be the optimal focal positions of the primary focus map in regions A, C, E, and G, respectively. Accordingly, in some embodiments, the primary focus map includes optimal focal positions corresponding to the depths of field for parts A2, B2, C2, D2, E2, F3, G3, H3, I3, and J3. In contrast to conventional approaches, the primary focus map, according to some embodiments in the disclosure, is generated based on images having the highest number of target cells in various regions of cytology specimen 500. Therefore, the primary focus map, according to some embodiments in the disclosure, corresponds to the presence of target cells.

[0026] Fig. 5B illustrates another exploded view of cytology specimen 500, arranged in accordance with some embodiments of the present disclosure. In conjunction with Fig. 2, in some embodiments, scanner 220 is configured to generate a secondary focus map across cytology specimen 500. More specifically, after generating the primary focus map, scanner 220 is further configured to compare images included in the same z-stack image and identify the image having target cells in the same z- stack image. For example, in region F, scanner 220 is configured to compare images included in the third z-stack image and identify the image associated with part F2 having target cells in the third z-stack image as illustrated in Fig. 2. Similarly, referring back to Fig. 5A, scanner 220 is configured to identify images of parts B1 , D1 , H1 , 11 , and J2 having target cells in the first, second, fourth, fifth, and sixth z- stack images, respectively.

[0027] In some embodiments, for non-scanned regions not scanned by scanner 220, scanner 220 is configured to estimate the optimal focal positions for the non-scanned regions (e.g., regions A, C, E, and G) based on the optical focal positions of the scanned regions (e.g., regions B, D, F, H, I, and J). The estimation may be based on some technical feasible approaches, such as linear or bilinear interpolation, triangulation, or more advanced computational methods to generate a smooth, continuous secondary focus map across the cytology specimen. Fig. 5B also illustrates an exploded view of cytology specimen 500 in which the depths of field forparts A1 , C1 , E1 , and G2 being estimated to be the optimal focal positions of the secondary focus map in regions A, C, E, and G, respectively. Accordingly, in some embodiments, the secondary focus map includes optimal focal positions corresponding to the depths of field for parts A1 , B1 , C1 , D1 , E1 , F2, G2, H1 , 11 , and J2.

[0028] In some embodiments, optimal focal positions associated with the secondary focus map are all either above or below their corresponding optimal focal positions associated with the primary focus map. In other words, the secondary focus map does not intersect with the primary focus map. For illustration, in conjunction with Fig. 5A and 5B, optical focal positions associated with the secondary focus map (i.e., the depths of field for parts A1 , B1 , C1 , D1 , E1 , F2, G2, H1 , 11 , and J2) are all above their corresponding optical focal positions associated with the primary focus map (i.e., the depths of field for parts A2, B2, C2, D2, E2, F3, G3, H3, I3, and J3). Therefore, in generating the secondary focus map, scanner 220 is configured to identify parts that are all either above or below corresponding parts associated with the primary focus map.

[0029] Compared to conventional approaches, the secondary focus map, according to some embodiments in the disclosure, is generated based on images having target cells in various regions of cytology specimen 500. Therefore, the secondary focus map, according to some embodiments in the disclosure, also corresponds to the presence of target cells. In addition, unlike a fixed distance between any corresponding focal position in the primary and secondary focus maps in the conventional approaches, a distance between any corresponding focal position in the primary and secondary focus maps varies across the cytology specimen 500 based on the distributions of target cells. For example, in conjunction with Fig. 5A and Fig. 5B, the focal positions in the secondary focus map are one layer above the focal positions in the primary focus map in regions A, B, C, D, E, F, G, and J but are two layers above the primary focus map in regions H and I. In other words, the distance between the target cells associated with the primary focus map and the target cells associated with the secondary focus map is different at corresponding focal positions in the primary and secondary focus maps.

[0030] Fig. 6 illustrates an example scanner 600 to generate one or more focus maps across a cytology specimen and obtain images associated with target cells in the cytology specimen based on the focus maps, arranged in accordance with some embodiments of the present disclosure. In conjunction with Fig. 2, in some embodiments, scanner 600 corresponds to scanner 220. Scanner 600 includes, but not limited to, computing device 610, camera 620, object lens module 630, stage 650 and light source 660. In some embodiments, stage 650 is configured to carry and move a glass slide. Cytology specimen 640 is distributed on the glass slide.

[0031] In some embodiments, computing device 610 includes a processor, a memory subsystem, and a communication subsystem. Artificial intelligence engines discussed above may be implemented as a set of executable instructions stored in the memory subsystem to be executed by the processor.

[0032] In some embodiments, computing device 610 is configured to generate control signals and transmit the control signals to camera 620, object lens module 630, stage 650, and light source 660 via the communication subsystem. For example, computing device 610 is configured to control light source 660 to generate lights with a specific range of wavelengths associated with image characteristics of target cells distributed in cytology specimen 640. For example, the specific range of wavelengths may correspond to the color information associated with the target cells as set forth above. An example specific range of wavelengths may be about 530 nm to about 630 nm. Alternatively, another example specific range of wavelengths may be about 450 nm to about 560 nm, and preferably about 450 nm to about 530 nm.

[0033] In some embodiments, computing device 610 is configured to control the movement of stage 650. When stage 650 carries the glass slide and cytology specimen 640, the movement of stage 650 may align a region (e.g., region A, B, C, D, E, F, G, H, I or J illustrated in Fig. 2) of cytology specimen 640 and a field of view of object lens 633 with a light path of the light generated by light source 660 and illustrated in Fig. 6. Therefore, the light generated by light source 660 may pass through the region and object lens 633 and reach camera 620.

[0034] In some embodiments, computing device 610 is configured to control lens switching module 631 of object lens module 630 to switch between different object lens (e.g., object lens 633 and 635). In some other embodiments, computing device 610 is configured to control camera 620, object lens module 630, and / or stage 650 to move so that object lens 633 or object lens 635 focuses on parts of cytology specimen 640. In conjunction with Fig. 2, object lens 633 may correspond to first object lens 220.

[0035] In some embodiments, computing device 610 is configured to control camera 620 to obtain an image of a focused part of cytology specimen 640. Computing device 610 is also configured to control camera 620 to send the obtained image to computing device 610 for further processing. Such processing includes, but not limited to, identifying images associated with target cells or non-target cells in cytology specimen 640 and comparing a z-stack image associated with the same region of cytology specimen 640 to determine the number of target cells in images of the z-stack image. In some embodiments, computing device 610 is configured to generate one or more focus maps across cytology specimen 640 based on the number of target cells captured in images of the z-stack image.

[0036] Fig. 7 is a flow diagram illustrating an example process 700 to obtain images associated with target cells distributed in a cytology specimen, arranged in accordance with some embodiments of the present disclosure. Process 700 may include one or more operations, functions, or actions as illustrated by blocks 710, 720, 730, 740, 750, 760, and / or 770 which may be performed by hardware, software and / or firmware. The various blocks are not intended to be limiting to the described embodiments. The outlined steps and operations are only provided as examples, and some of the steps and operations may be optional, combined into fewer steps and operations, or expanded into additional steps and operations without detracting from the essence of the disclosed embodiments. Although the blocks are illustrated in a sequential order, these blocks may also be performed in parallel, and / or in a different order than those described herein. In some embodiments, process 700 may be applied to various scanning approaches to scan a cytology specimen withcorresponding different whole-slide imaging scanners. Such scanning approaches may include, but not limited to, area scanning or line scanning approaches.

[0037] Process 700 may begin at block 710, “obtain first z-stack image including first image and second image.” In some embodiments, in conjunction with Fig. 2 and Fig. 5A, scanner 220 is configured to obtain a first z-stack image associated with a first region of cytology specimen 500 in a focusing process over the first region. In some embodiments, in the focusing process, scanner 220 is configured to focus on target cells distributed in the first region. In some embodiments, scanner 220 is configured to randomly select and scan the first region (e.g., region B) of cytology specimen 500 among all regions of cytology specimen 500. In some other embodiments, scanner 220 is configured to prioritize selecting and scanning the first region of cytology specimen 500 based on image information (e.g., indicating some regions are occupied while other regions are empty) obtained in an earlier scan of cytology specimen 500 by scanner 220 (e.g., a lower magnification scan of scanner 220). Example image information may include, but not limited to, an intensity value of an image obtained in the earlier scan (e.g., indicating pixels of the image to be nonwhite or less white) and expected size, structure, and appearance of cells. After selecting region B, scanner 220 is further configured to obtain a first z-stack image associated with region B. For illustration, scanner 220 is configured to obtain a first image associated with part B2 and a second image associated with part B1 . In some embodiments, the first image is associated with a first set of target cells distributed in region B, and the second image is associated with a second set of target cells distributed in region B. In some embodiments, a number of the first set of target cells corresponds to the highest number of target cells distributed in the parts corresponding to region B.

[0038] In some embodiments, scanner 220 is configured to determine whether a number of images included in the first z-stack image reaches a threshold. For illustration, suppose a threshold of four images is set for the first z-stack image. Scanner 220 is configured to further obtain a third image associated with part B3 and a fourth image associated with part B4 to reach the threshold of four.

[0039] Block 710 may be followed by block 720, “obtain second z-stack image including third image and fourth image.” In some embodiments, in conjunction with Fig. 2 and Fig. 5A, scanner 220 is configured to obtain a second z-stack image associated with a second region of cytology specimen 500 in a focusing process over the second region. In some embodiments, in the focusing process, scanner 220 is configured to focus on target cells distributed in the parts corresponding to the second region. In some embodiments, scanner 220 is configured to randomly select the second region (e.g., region I) of cytology specimen 500. After selecting region I, scanner 220 is further configured to obtain a second z-stack image associated with region I. For illustration, scanner 220 is configured to obtain a third image associated with part I3 and a second image associated with part 11 . In some embodiments, the third image is associated with a third set of target cells distributed in region I, and the fourth image is associated with a fourth set of target cells distributed in region I. In some embodiments, a number of the third set of target cells corresponds to the highest number of target cells distributed in region I.

[0040] In some embodiments, scanner 220 is configured to determine whether a number of images included in the second z-stack image reaches a threshold. For illustration, assuming a threshold of four images to be included in the second z-stack image, scanner 220 is configured to further obtain a fifth image associated with part I2 and a sixth image associated with part I4.

[0041] Block 720 may be followed by block 730, “generate first focus map.” In conjunction with Fig. 2 and Fig. 5A, scanner 220 is configured to generate a first focus map across cytology specimen 500. In some embodiments, the first focus map corresponds to the primary focus map discussed in the descriptions of Fig. 5A. In some embodiments, the first focus map includes part B2 and part I3.

[0042] In some embodiments, scanner 220 is configured to obtain z-stack images associated with one or more additional regions (e.g., region D) of cytology specimen 500. In some embodiments, in response to determining that a number of regions from which z-stack images are obtained does not reach a threshold number of regions, scanner 220 is configured to obtain z-stack images associated with other regions (e.g., regions F, H, and J). In some embodiments, based on the z-stackimages, the first focus map further includes parts D2, F3, H3, and J3 as illustrated in Fig. 5A.

[0043] Block 730 may be followed by block 740, “capture image according to first focus map.” In conjunction with Fig. 2 and Fig. 5A, scanner 220 is configured to capture images of target cells according to the first focus map including parts B2, D2, F3, H3, 13, and J3. In some embodiments, scanner 220 is also configured to capture images of target cells according to the first focus map including A2, C2, E2, and G3 through estimations.

[0044] Block 740 may be followed by block 750, “generate second focus map.” In conjunction with Fig. 2 and Fig. 5A, scanner 220 is configured to generate a second focus map across cytology specimen 500. In some embodiments, the second focus map corresponds to the secondary focus map discussed in the descriptions of Fig. 5B. In some embodiments, the second focus map includes part B1 and part 11 .

[0045] In some embodiments, the distance between target cells at any corresponding focal position in the first and second focus maps is varied across cytology specimen 500 based on the distributions of target cells. For illustration, a first distance between a set of target cells distributed in part B2 included in the first focus map and a set of target cells distributed in part B1 included in the second focus map is different from a second distance between a set of target cells distributed in part I3 included in the first focus map and a set of target cells distributed in part 11 included in the second focus map.

[0046] Block 750 may be followed by block 760, “number of captured target cells less than threshold.” In some embodiments, in conjunction with Fig. 2 and Fig. 5A, scanner 220 is configured to determine a number of target cells captured in block 740 less than a threshold number of target cells. In some embodiments, in response to determining the number of target cells captured in block 740 greater than or equal to the threshold number of target cells, process 700 may end and there is no need to further capture images of target cells, which can save scanning time.

[0047] In some other embodiments, in response to determining the number of target cells captured in block 740 less than the threshold number of target cells, block 760may be followed by block 770, “capture image according to second focus map.” In conjunction with Fig. 2 and Fig. 5B, scanner 220 is configured to capture images of target cells according to the second focus map including parts B1 , D1 , F2, H1 , 11 , and J2. In some embodiments, scanner 220 is also configured to capture images of target cells according to the second focus map including A1 , C1 , E1 , and G2 through estimations.

[0048] The above examples can be implemented by hardware (including hardware logic circuitry), software or firmware or a combination thereof. Fig. 8 is an example device 800 configured to perform various embodiments of the present disclosure. For example, device 800 can be implemented as a server configured to perform process 700 in Fig. 7. Device 800 can also be implemented as an intermediary network device configured to perform process 700 in Fig. 7. Device 800 may be any computing device or networking device suitable for practicing one or more embodiments of the present disclosure. It is noted that device 800 described herein is illustrative and that any other technically feasible configurations fall within the scope of the present disclosure.

[0049] As shown, device 800 includes, without limitation, an interconnect (bus) 830 that connects at least one processor 840, computer-readable medium 810, input / output (I / O) device interface 850 and network interface 860. Processor 840 may be any suitable processor implemented as a central processing unit (CPU), a graphics processing unit (GPU), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), any other type of processing unit, or a combination of different processing units, such as a CPU configured to operate in conjunction with a GPU or digital signal processor (DSP). In general, processor 840 may be any technically feasible hardware unit capable of processing data and / or executing executable instructions, including process 870. In some embodiments, process 870 may correspond to process 700 set forth above.

[0050] Network interface 860 is configured to couple device 800 to one or more networks, so that device 800 can communicate with other devices on such network(s).

[0051] Processor 840, I / O device interface 850 and network interface 860 are configured to read data from and write data to computer-readable medium 810. Computer-readable medium 810 may have stored thereon instructions or program code that, in response to execution by the processor, cause the processor to perform the process described herein with reference to Fig. 7. In some embodiments, computer-readable medium 810 includes cache 820, which can be used to store data associated with process 870.

[0052] Some aspects of the embodiments disclosed herein, in whole or in part, can be equivalently implemented in integrated circuits, as one or more programs running on one or more processors, as firmware, or as virtually any combination thereof, and that designing the circuitry and / or writing the code for the software and or firmware are possible in light of this disclosure.

[0053] Software and / or other instructions to implement the techniques introduced here may be stored on a non-transitory computer-readable storage medium (e.g., 810) and may be executed by one or more special-purpose programmable microprocessors, such as processor 840. A “computer-readable storage medium”, as the term is used herein, includes any mechanism that provides (i.e., stores and / or transmits) information in a form accessible by a machine (e.g., a computer, network device, personal digital assistant (PDA), mobile device, manufacturing tool, any device with a set of one or more processors, etc.). A computer-readable storage medium may include recordable / non-recordable media (e.g., read-only memory (ROM), random access memory (RAM), magnetic disk or optical storage media, flash memory devices, solid state storage, etc.)

[0054] From the foregoing, it will be appreciated that various embodiments of the present disclosure have been described herein for illustration purposes, and that various modifications may be made without departing from the scope and spirit of the present disclosure. Accordingly, the various embodiments disclosed herein are not intended to be limiting.

Claims

CLAIMS:1 . A method to capture images of target cells distributed in a cytology specimen, comprising: obtaining a first z-stack image associated with a first region of the cytology specimen through a lens, wherein the first z-stack image includes a first image obtained at a first depth of field from a surface of the cytology specimen and a second image obtained at a second depth of field from the surface of the cytology specimen, and the first image is associated with a first set of target cells distributed in the first region and the second image is associated with a second set of target cells distributed in the first region; obtaining a second z-stack image associated with a second region of the cytology specimen through the lens, wherein the second z-stack image includes a third image obtained at a third depth of field from the surface of the cytology specimen and a fourth image obtained at a fourth depth of field from the surface of the cytology specimen, and the third image is associated with a third set of target cells distributed in the second region and the fourth image is associated with a fourth set of target cells distributed in the second region; generating a first focus map based on the first region, the second region, the first depth of field, and the third depth of field; capturing images of the target cells according to the first focus map through the lens; generating a second focus map based on the first region, the second region, the second depth of field, and the fourth depth of field, wherein a first distance along a z-axis of the cytology specimen between the first set of target cells and the second set of target cells is different from a second distance along the z-axis of the cytology specimen between the third set of target cells and the fourth set of target cells; and in response to a number of the target cells obtained according to the first focus map less than a first threshold, capturing images of the target cells according to the second focus map through the lens.

2. The method of claim 1 , wherein a number of the first set of target cells corresponds to the highest number of target cells distributed in the first region, and a number of the third set of target cells corresponds to the highest number of target cells distributed in the second region.

3. The method of claim 1 , wherein the second focus map is above the first focus map in response to the second depth of field being less than the first depth of field and the fourth depth of field being less than the third depth of field.

4. The method of claim 1 , wherein the second focus map is below the first focus map in response to the second depth of field being greater than the first depth of field and the fourth depth of field being greater than the third depth of field.

5. The method of claim 1 , further comprising: determining whether a number of images included in the first z-stack image reaches a second threshold; adjusting a distance between a stage configured to hold the cytology specimen and the lens in response to determining the number of images does not reach the second threshold; and obtaining additional images associated with the first region through the lens.

6. The method of claim 1 , wherein the first z-stack image is obtained in a first focusing process of the lens over the first region and the second z-stack image is obtained in a second focusing process of the lens over the second region.

7. The method of claim 6, wherein the lens is configured to focus on the target cells in the first region in the first focusing process and on the target cells in the second region in the second focusing process.

8. The method of claim 1 , further comprising obtaining z-stack images associated with additional regions of the cytology specimen through the lens.

9. The method of claim 8, further comprising:determining whether a number of regions including the first region, the second region, and the additional regions reaches a third threshold; and obtaining z-stack images associated with other regions of the cytology specimen slide through the lens in response to determining the number of regions does not reach the third threshold.

10. The method of claim 1 , further comprising prioritizing the first region among all regions of the cytology specimen to obtain the first z-stack image based on image information obtained in an earlier scan of the cytology specimen.11 . A system to obtain images of target cells distributed in a cytology specimen, comprising: a lens; a processor; and a non-transitory computer-readable medium having stored thereon instructions that, when executed by the processor, cause the processor to: obtain a first z-stack image associated with a first region of the cytology specimen through the lens, wherein the first z-stack image includes a first image obtained at a first depth of field from a surface of the cytology specimen and a second image obtained at a second depth of field from the surface of the cytology specimen, and the first image is associated with a first set of target cells distributed in the first region and the second image is associated with a second set of target cells distributed in the first region; obtain a second z-stack image associated with a second region of the cytology specimen through the lens, wherein the second z-stack image includes a third image obtained at a third depth of field from the surface of the cytology specimen and a fourth image obtained at a fourth depth of field from the surface of the cytology specimen, and the third image is associated with a third set of target cells distributed in the second region and the fourth image is associated with a fourth set of target cells distributed in the second region; generate a first focus map based on the first region, the second region, the first depth of field, and the third depth of field;obtain images of the target cells according to the first focus map; generate a second focus map based on the first region, the second region, the second depth of field, and the fourth depth of field, wherein a first distance along a z-axis of the cytology specimen between the first set of target cells and the second set of target cells is different from a second distance along the z-axis of the cytology specimen between the third set of target cells and the fourth set of target cells; and in response to a number of the target cells obtained according to the first focus map less than a first threshold, obtain images of the target cells according to the second focus map.

12. The system of claim 11 , wherein a number of the first set of target cells corresponds to the highest number of target cells distributed in the first region, and a number of the third set of target cells corresponds to the highest number of target cells distributed in the second region.

13. The system of claim 11 , wherein the second focus map is above the first focus map in response to the second depth of field being less than the first depth of field and the fourth depth of field being less than the third depth of field.

14. The system of claim 11 , wherein the second focus map is below the first focus map in response to the second depth of field being greater than the first depth of field and the fourth depth of field being greater than the third depth of field.

15. The system of claim 11 , wherein the non-transitory computer-readable medium having stored thereon additional instructions that, when executed by the processor, cause the processor to: determine whether a number of images included in the first z-stack image reaches a second threshold; adjust a distance between a stage configured to hold the cytology specimen and the lens in response to determining the number of images does not reach the second threshold; and obtain additional images associated with the first region through the lens.

16. The system of claim 11 , wherein the first z-stack image is obtained in a first focusing process of the lens over the first region and the second z-stack image is obtained in a second focusing process of the lens over the second region.

17. The system of claim 16, wherein the lens is configured to focus on the target cells in the first region in the first focusing process and on the target cells in the second region in the second focusing process.

18. The system of claim 11 , wherein the non-transitory computer-readable medium having stored thereon additional instructions that, when executed by the processor, cause the processor to obtain z-stack images associated with additional regions of the cytology specimen through the lens.

19. The system of claim 18, wherein the non-transitory computer-readable medium having stored thereon additional instructions that, when executed by the processor, cause the processor to: determine whether a number of regions including the first region, the second region, and the additional regions reaches a third threshold; and obtain z-stack images associated with other regions of the cytology specimen slide through the lens in response to determining the number of regions does not reach the third threshold.

20. A non-transitory computer-readable storage medium that includes a set of instructions which, in response to execution by a processor of a computing device, causes the processor to perform a method to capture images of target cells distributed in a cytology specimen, the method comprising: obtaining a first z-stack image associated with a first region of the cytology specimen through the lens, wherein the first z-stack image includes a first image obtained at a first depth of field from a surface of the cytology specimen and a second image obtained at a second depth of field from the surface of the cytology specimen, and the first image is associated with afirst set of target cells distributed in the first region and the second image is associated with a second set of target cells distributed in the first region; obtaining a second z-stack image associated with a second region of the cytology specimen through the lens, wherein the second z-stack image includes a third image obtained at a third depth of field from the surface of the cytology specimen and a fourth image obtained at a fourth depth of field from the surface of the cytology specimen, and the third image is associated with a third set of target cells distributed in the second region and the fourth image is associated with a fourth set of target cells distributed in the second region; generating a first focus map based on the first region, the second region, the first depth of field, and the third depth of field; capturing images of the target cells according to the first focus map; generating a second focus map based on the first region, the second region, the second depth of field, and the fourth depth of field, wherein a first distance along a z-axis of the cytology specimen between the first set of target cells and the second set of target cells is different from a second distance along the z-axis of the cytology specimen between the third set of target cells and the fourth set of target cells; and in response to a number of the target cells captured according to the first focus map less than a first threshold, capturing images of the target cells according to the second focus map.

Images