Back to results
US 12,569,128 B2Granted
Apparatus and method for detecting cervical cancer
AssigneeBoard of Regents of the Nevada System of Higher Education, on behalf of the University of Nevada, Reno
InventorsSesh Commuri, Garrett Winkelmaier, Paloma Cepeda Andrade
Life of this patent
Estimating the term…
Abstract
A method for automatic, image-based detection of abnormalities in a patient's cervix is disclosed. In an embodiment, a Cervitude Imaging System (CIS) processor of a CIS device illuminates a light source affixed to a distal end of a probe housing, receives cervical image data of the patient from a camera affixed to the distal end of the probe housing, generates a reconstructed image that includes reduced specular reflections, and segments the reconstructed image into at least one region of interest (ROI). The process also includes the CIS processor transmitting the reconstructed image comprising the at least one ROI to at least one of a host system and an input/output device.
Claims16
A method for automatic, image-based detection of abnormalities in a patient's cervix comprising:
- illuminating, by a processor, a light source affixed to a distal end of a probe housing;
- receiving, by the processor, cervical image of the patient from a camera affixed to the distal end of the probe housing, wherein the cervical image comprises cervical image data;
- receiving, by the processor, ranging data from a ranging sensor affixed to the distal end of the probe housing;
- generating, by the processor based on the cervical image data, a reconstructed image having reduced specular reflections, wherein the generating a reconstructed image comprises:
- generating a brightness map using the cervical image data, wherein the generating a brightness map comprises:
- converting cervical image data to a CIELAB color space, wherein the CIELAB color space comprises hue values, a saturation values, and a luminosity values, wherein the luminosity values are in the range of 0 to 100 where 0 is black and 100 is white;
- generating a structuring element, wherein the structuring element comprises a binary matrix or a binary image, and wherein the structuring element can be added to the cervical image data to make a filtered image;
- performing erosion on the cervical image, wherein the performing erosion comprises:
- removing any foreground structures that are smaller than the structuring element, wherein the removing any foreground structures comprises:
- adding the structuring element to the cervical image data at one or more positions of the cervical image;
- and
- deleting any foreground pixels of a cervical image of the cervical image data that are not fully contained by the structuring element;
- performing dilation on the cervical image, wherein the performing dilation comprises:
- generating a point set, wherein the generating a point set comprises a vector sum of one or more pairs of coordinate points from the cervical image data and the structuring element, wherein the coordinate points comprise foreground pixels and background pixels, and wherein the coordinate pairs of all foreground pixels have a value of 1;
- and
- discarding areas of the cervical image where a specular reflection is predominant to make a brightness map, wherein the discarding areas of the cervical image comprises performing erosion on the cervical image followed by performing dilation using the same structuring element to remove one or more foreground structures that are smaller than the structuring element and the remaining structures are smoothed by the ensuing dilation;
- and
- normalizing the luminosity values of the brightness map to make the reconstructed image;
- segmenting, by the processor, the reconstructed image into at least one region of interest (ROI) of the cervix;
- identifying, by the processor, a lesion in the ROI of the cervix;
- determining, by the processor based on the ranging data, a distance of the cervix from the camera and a view angle of the camera;
- and
- generating, by the processor based on the distance of the cervix from the camera, a size of the lesion.
20 more elements
A portable cervical imaging device for capturing cervical images of a patient comprising:
- a probe housing comprising a camera, a ranging sensor, and input/output device and a light source;
- and
- a control housing connected to the probe housing, the control housing comprising a CIS processor, a communication device, a power supply and a storage device, wherein the storage device stores processor executable instructions which when executed causes the processor to:
- illuminate the light source;
- receive from the camera cervical image data of the patient, wherein the cervical image data comprises Red Green Blue (RGB) color space with values in the range from 0 to 255 for each color;
- receive ranging data from the ranging sensor;
- generate a reconstructed image based on the cervical image data having reduced specular reflections, wherein the generating a reconstructed image comprises:
- generating a brightness map using the cervical image data, wherein the generating a brightness map comprises:
- converting cervical image data to a CIELAB color space, wherein the CIELAB color space comprises hue values, a saturation values, and a luminosity values, wherein the luminosity values are in the range of 0 to 100 where 0 is black and 100 is white;
- generating a structuring element, wherein the structuring element comprises a binary matrix or a binary image, and wherein the structuring element can be added to the cervical image data to make a filtered image;
- performing erosion on the cervical image, wherein the performing erosion comprises:
- removing any foreground structures that are smaller than the structuring element, wherein the removing any foreground structures comprises:
- adding the structuring element to the cervical image data at one or more positions of the cervical image;
- and
- deleting any foreground pixels of a cervical image of the cervical image data that are not fully contained by the structuring element;
- performing dilation on the cervical image, wherein the performing dilation comprises:
- generating a point set, wherein the generating a point set comprises a vector sum of one or more pairs of coordinate points from the cervical image data and the structuring element, wherein the coordinate points comprise foreground pixels and background pixels, and wherein the coordinate pairs of all foreground pixels have a value of 1;
- and
- discarding areas of the cervical image where a specular reflection is predominant to make a brightness map, wherein the discarding areas of the cervical image comprises performing erosion on the cervical image followed by performing dilation using the same structuring element to remove one or more foreground structures that are smaller than the structuring element and the remaining structures are smoothed by the ensuing dilation;
- and
- normalizing the luminosity values of the brightness map to make the reconstructed image;
- segment the reconstructed image into at least one region of interest (ROI) of the cervix;
- identify a lesion in the ROI of the cervix;
- determine, based on the ranging data, a distance of the cervix from the camera and a view angle of the camera;
- generate, based on the distance of the cervix from the camera, a size of the lesion.
22 more elements
Dependencies are inferred from the claim text.