Computer vision-based monitoring system for jaundice treatment

Abstract

The present invention relates to a computer-based jaundice monitoring system that uses image processing technology to monitor and detect jaundice in a patient. The system includes a camera for capturing patient images, a computing device for image processing and analysis, a local communication network for interconnecting system components, and a light panel integrated with the camera for providing phototherapy. The computing device identifies a patient's skin site, analyzes the dominant color within the skin site, detects yellow pigmentation indicative of jaundice, and generates a notification accordingly.

Description

Technical Field

[0001] The present invention relates to a patient monitoring system. More particularly, the present invention relates to a computer vision-based monitoring system for jaundice treatment.

Background Art

[0002] Neonatal jaundice is a condition commonly seen in newborns where the skin and eyes turn yellow due to an excess of bilirubin. Recent studies have shown that it is widespread, particularly in low- and middle-income countries. Risk factors such as premature infants and low birth weight infants have been identified. Early detection and early treatment are important to prevent severe cases. Phototherapy using blue light is the standard treatment method, but conventional systems have limitations. Phototherapy using LEDs is more cost-effective and efficient. Computer vision technology has emerged as a promising solution for improving jaundice management, especially in resource-limited environments. By using computer vision algorithms to analyze skin color, doctors and healthcare workers can accurately diagnose and monitor jaundice without relying on expensive equipment or specialized training. This technology also enables the development of low-cost, portable lights for phototherapy that can be easily introduced even in areas with limited access to healthcare. The present invention aims to contribute to the improvement of neonatal medical solutions by providing an inexpensive, readily available, and effective phototherapy device that can be used even in low-resource environments. By combining phototherapy treatment and non-invasive skin color monitoring, the present invention has the potential to improve the accuracy and effectiveness of jaundice treatment and reduce the risks and complications associated with untreated jaundice.

[0003] Currently, various patient monitoring systems have been disclosed in this field. For example, Hashim et al. (2021) disclose a neonatal jaundice detection system that detects skin areas by converting images to the Lab color space to obtain the values ​​of the B and Cb channels and comparing these values ​​to a predefined threshold. If the value of either the B or Cb channel is smaller than the threshold, the diagnosis is "jaundice," and the infant requires phototherapy.

[0004] Otherwise, the diagnosis is "normal".

[0005] Patent Document 1 discloses a newborn monitoring system. This system consists of a camera equipped with a face recognition module and a body recognition module. The system also includes a face color recognition module that can identify jaundice and activates an emergency module when jaundice is detected.

[0006] However, these patient monitoring systems do not disclose a patient monitoring system as described in the present invention. [Prior art documents] [Patent Documents]

[0007] [Patent Document 1] Korean Patent Publication No. 10-1623167 [Overview of the project] [Problems that the invention aims to solve]

[0008] The objective of this invention is to provide a system for monitoring jaundice. [Means for solving the problem]

[0009] Therefore, the present invention provides a computer vision-based jaundice monitoring system. The system is a. A camera configured to capture images of a patient, b. A computing device programmed to communicate with a camera and process captured images, wherein the processing is: i. A step of identifying the patient's skin area in the captured image, ii. A step of analyzing the dominant color present in the patient's skin area, iii. A step to determine the presence or absence of yellow pigmentation based on the dominant color, iv. A computing device comprising the step of generating and transmitting a notification when yellow pigmentation is detected, c. A light panel integrated with a camera for administering phototherapy to a patient, d. A local communication network that facilitates connection between the camera, the computing unit, and the optical panel. [Brief explanation of the drawing]

[0010] [Figure 1] Figure 1 illustrates a computer vision-based jaundice monitoring system according to a preferred embodiment of the present invention. [Figure 2] Figure 2 illustrates a method for monitoring jaundice using a computer-based system according to a preferred embodiment of the present invention. [Modes for carrying out the invention]

[0011] Figure 1 illustrates a computer vision-based jaundice monitoring system according to a preferred embodiment of the present invention. The system comprises a camera 100 configured to capture images of a patient. A computing device 102 communicating with the camera 100 processes the captured images. The image processing consists of identifying areas of the patient's skin within the captured images.

[0012] Next, the computing unit 102 analyzes the dominant color present in the patient's skin area. Then, it determines whether or not there is yellow pigmentation based on the dominant color. Finally, if yellow pigmentation is determined to be present, the computing unit 102 generates and transmits a notification.

[0013] The system further includes a light panel 104 provided integrally with the camera 100. The light panel 104 provides phototherapy to the patient.

[0014] Furthermore, the system includes a local communication network 106, which enables easy connection among the camera 100, the computing device 102, and the light panel 104. Preferably, the local communication network 106 is a Wi-Fi network. Alternatively, the local communication network 106 is a Bluetooth (registered trademark) network or other communication network technologies available in the art.

[0015] In a preferred embodiment, the computing device 102 transmits a signal that controls the intensity of the light panel 104 based on the yellow pigmentation of the patient's skin.

[0016] In another further embodiment, a monitoring device 108 is further connected to the local communication network 106. The monitoring device 108 receives notifications and data from the computing device 102.

[0017] FIG. 2 is a diagram illustrating a method for monitoring jaundice using a computer-based system according to a preferred embodiment of the present invention. This method starts with the step of capturing an image of the patient using a camera (step 200). Next, a computing device communicating with the camera processes the captured image (step 202). This processing includes the step of identifying the patient's skin area within the captured image (step 204). Preferably, the identification is performed by identifying pixels as skin using the following rules: (i) R>95 and G>40 and B>20(1) (ii) Max(R,G,B)-Min(R,G,B)>15 (iii) |R-G|>15 (iv) R>G and R>B Here, R is the red channel, G is the green channel, and B is the blue channel.

[0018] The first rule defines the range of RGB values, the second rule describes the range of saturation, the third rule specifies the acceptable difference between the red and green channels, and the fourth rule requires that the red channel be greater than the green and blue channels. When all these rules are satisfied, the pixel is considered a skin pixel.

[0019] Next, the computing device analyzes the dominant color within the patient's skin area (step 206). This is done by k-means clustering.

[0020] Next, the computing device determines whether the dominant color indicates yellow pigmentation (step 208). The b* value of the color in the CIE L*a*b color space is used as the key variable. CIE L*a*b* is a device-independent 3D color space that enables accurate measurement and comparison of all perceptible colors represented by three values: L* is the perceptual lightness, and a* and b* indicate chromaticity coordinates.

[0021] The yellow score is determined by the weighted average of the normalized b* coordinates of the extracted dominant skin color with respect to a reference yellow point and a reference "normal" skin point. This is to provide a more interpretable data representation for monitoring purposes. The calculation formula is as follows: Yellow score = Σiwi(bi - bmin) / (bmax - bmin) Here, - bi is the b* coordinate of the position of the extracted color i in the CIE L*a*b* color space; - bmin is the b* coordinate of a reference normal skin pixel; - bmax is the b* coordinate of the reference maximum yellow point; - wi is the weight assigned to each extracted color i, determined by the number of pixels gathered around it (how dominant it is).

[0022] After that, the computing device generates a notification upon detection of yellow pigmentation (step 210). The computing unit also transmits signals to control the intensity of the light panel, thereby administering phototherapy to the patient (step 212).

[0023] In a further embodiment of the present invention, the computing device utilizes a machine learning algorithm to improve the accuracy of determining yellow pigmentation.

[0024] The embodiments described herein are intended to be extended to the individual elements and concepts described herein, independently of any other concepts, ideas, or systems, and to include combinations of elements described elsewhere in this application. It should be understood that the present invention is not limited to the embodiments described in detail herein with reference to the accompanying drawings. Thus, many embodiments and variations will be apparent to those skilled in the art. The exemplary embodiments described herein represent preferred forms, but are not limited thereto, and modifications and alternative forms are possible. Accordingly, the scope of the invention is intended to be defined by the following claims and their equivalents. Furthermore, elements described individually or as part of an embodiment are intended to be combined with other individually described elements or parts of other embodiments, even if other elements and embodiments do not refer to such elements. Therefore, the absence of a combination described herein does not prevent the inventor from claiming rights to such combination.

Claims

1. A computer vision-based jaundice monitoring system, a. A camera configured to capture images of a patient, b. A computing device programmed to communicate with the camera and process the captured image, wherein the processing is: i. The step of identifying the skin area of ​​the patient in the captured image, ii. A step of analyzing the dominant color present in the skin area of ​​the patient by k-means clustering, iii. A step of determining whether the dominant color exhibits yellow pigmentation by using the b* value of the color in the CIE L*a*b* color space as a key variable, wherein L* represents perceptual lightness and a* and b* represent chromaticity coordinates, iv. A computing device comprising the step of generating and transmitting a notification when yellow pigmentation is detected, c. A light panel integrated with the camera for administering phototherapy to the patient, d. A local communication network that easily connects the camera, the computing device, and the optical panel, e. A monitoring device connected to the local communication network and receiving the notification and patient data, Equipped with, The yellow score is determined by a weighted average of the normalized b* coordinates of the extracted dominant color relative to a reference yellow point and a reference "normal" skin point: Yellow score = Σiwi(bi - bmin) / (bmax - bmin) Here, -bi is the b* coordinate of the position of the extracted color i in the CIE L*a*b* color space; -bmin is the b* coordinate of the reference normal skin pixel; -bmax is the b* coordinate of the reference yellow maximum point; -wi is a computer vision-based jaundice monitoring system in which each extracted color i is assigned a weight determined by the number of pixels clustered around it.

2. The system according to claim 1, wherein the computing device utilizes a machine learning algorithm to improve the accuracy of determining yellow pigmentation.

3. The system according to claim 1, wherein the local communication network is a wireless network utilizing Wi-Fi or Bluetooth® technology.

4. The system according to claim 1, wherein the light panel is adjustable in intensity to control and provide phototherapy to the patient.

5. A method for monitoring jaundice using a computer-based system, a. The step of taking an image of the patient using a camera, b. A step of processing the captured image using a computing device that communicates with the camera, i. A step of identifying the skin area of ​​the patient in the captured image, ii. The step of analyzing the dominant color within the skin area of ​​the patient by k-means clustering, iii. A step of determining whether the dominant color exhibits yellow pigmentation indicative of jaundice, using the b* value of the color in the CIE L*a*b* color space as a key variable, wherein L* represents perceptual lightness and a* and b* represent chromaticity coordinates, iv. A step of processing the image, which includes the step of generating a notification when yellow pigmentation is detected and transmitting the notification and patient data to a monitoring device connected to a local communication network connecting the camera and the computing device, Includes, The yellow score is determined by a weighted average of the normalized b* coordinates of the extracted dominant color relative to a reference yellow point and a reference "normal" skin point: Yellow score = Σiwi(bi - bmin) / (bmax - bmin) Here, -bi is the b* coordinate of the position of the extracted color i in the CIE L*a*b* color space; -bmin is the b* coordinate of the reference normal skin pixel; -bmax is the b* coordinate of the reference yellow maximum point; -wi is a method for monitoring jaundice, where each extracted color i is assigned a weight determined by the number of pixels clustered around it.

Images