CN-122023251-A - Lesion area identification method and device, electronic equipment and storage medium

CN122023251ACN 122023251 ACN122023251 ACN 122023251ACN-122023251-A

Abstract

The application relates to the technical field of medical image processing, and discloses a lesion area identification method, a lesion area identification device, electronic equipment and a storage medium. The method comprises the steps of obtaining a medical image of any human body component part, dividing and identifying lesions in the medical image to obtain a plurality of lesion areas and at least one transition area, selecting any lesion area as a target lesion area, taking each lesion area adjacent to the target lesion area as a to-be-determined lesion area to determine a reference distance between the target lesion area and each to-be-determined lesion area, combining the target lesion area and any transition area to obtain combined lesion areas, determining the actual distance between each combined lesion area and the corresponding to-be-determined lesion area, and determining whether the corresponding lesion area and transition area belong to the same real lesion area or not based on the reference distance and the actual distance. The application can at least improve the identification degree of the lesion boundary and is beneficial to reducing the risk of misdiagnosis.

Inventors

Mao Guoqun
WEI FUQUAN
XU WENJIE
HUANG YIJIANG
CHEN NAN
LV LEI
HUANG YIMIN

Assignees

浙江省立同德医院(浙江省精神卫生研究院)

Dates

Publication Date: 20260512
Application Date: 20251225

Claims (10)

1. A lesion area identification method, comprising at least: Acquiring a medical image of any human body component, and dividing and identifying lesions in the medical image to obtain a plurality of lesion areas and at least one transition area; selecting any lesion area as a target lesion area, and taking each lesion area adjacent to the target lesion area as a pending lesion area to determine a reference distance between the target lesion area and each pending lesion area; Combining the target lesion areas with any transition area to obtain combined lesion areas, determining the actual distance between each combined lesion area and the corresponding undetermined lesion area, and further determining whether the corresponding target lesion areas, transition areas and undetermined lesion areas belong to the same real lesion area or not based on the reference distance and the actual distance.
2. The lesion field identification method according to claim 1, wherein after said combining the target lesion field with any one of the transition fields to obtain combined lesion fields, determining an actual distance between each of the combined lesion fields and the corresponding pending lesion field, and further determining whether the corresponding target lesion field, the transition field, and the pending lesion field belong to the same real lesion field based on the reference distance and the actual distance, at least further comprising: when the corresponding target lesion area, the transition area and the undetermined lesion area belong to the same real lesion area, determining at least lesion volume and uniformity according to the real lesion area; and acquiring at least one lesion sign index except the lesion volume and the uniformity, and judging the degree of deterioration of the lesion according to at least one lesion sign index, the lesion volume and the uniformity.
3. The lesion field identification method according to claim 1, wherein the segmentation identification of lesions in the medical image is achieved at least by a preset lesion segmentation model configured with confidence values; The confidence value corresponding to the background area in the medical image is at least a first threshold value; the confidence value corresponding to the lesion area in the medical image is at least not lower than a second threshold value; The confidence value corresponding to the transition region in the medical image is at least between a third threshold and a fourth threshold.
4. The lesion area identification method according to claim 3, wherein the first threshold value is less than the third threshold value, the third threshold value is less than the fourth threshold value, and the fourth threshold value is not greater than the second threshold value.
5. The lesion field identification method according to claim 1, wherein the reference distance is at least the shortest distance between the target lesion field and each of the pending lesions fields.
6. The lesion area identification method according to claim 1, wherein when the actual distance is smaller than the reference distance, it is determined that the corresponding target lesion area, transition area, and pending lesion area belong to the same real lesion area; and when the actual distance is not smaller than the reference distance, determining that the corresponding target lesion region, the transition region and the undetermined lesion region do not belong to the same real lesion region.
7. A lesion area identification method, comprising at least: Acquiring a medical image of any human body component, and dividing and identifying lesions in the medical image to obtain a plurality of lesion areas and at least one transition area; selecting any lesion area as a target lesion area, and taking each lesion area adjacent to the target lesion area as a pending lesion area to determine a reference distance between the target lesion area and each pending lesion area; and combining any undetermined lesion area with any transition area to obtain combined lesion areas, determining the actual distance between each combined lesion area and the target lesion area, and further determining whether the corresponding target lesion area, transition area and undetermined lesion area belong to the same real lesion area or not based on the reference distance and the actual distance.
8. A lesion area identification device for performing the lesion area identification method according to any one of claims 1-6; The lesion area identifying device includes at least: The region segmentation module is used for acquiring a medical image of any human body component part, and carrying out segmentation recognition on lesions in the medical image so as to obtain a plurality of lesion regions and at least one transition region; The distance determining module is used for selecting any lesion area as a target lesion area, and taking each lesion area adjacent to the target lesion area as a to-be-determined lesion area so as to determine a reference distance between the target lesion area and each to-be-determined lesion area; The attribution judging module is used for combining the target lesion area and any transition area to obtain combined lesion areas, determining the actual distance between each combined lesion area and the corresponding undetermined lesion area, and further determining whether the corresponding target lesion area, transition area and undetermined lesion area belong to the same real lesion area or not based on the reference distance and the actual distance.
9. An electronic device comprising a memory and a processor, the memory storing a computer program executable on the processor, characterized in that the processor implements the steps of the lesion area identification method according to any one of claims 1 to 7 when the program is executed.
10. A computer-readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps in the lesion area identification method according to any one of claims 1 to 7.

Description

Lesion area identification method and device, electronic equipment and storage medium Technical Field The present invention relates to the field of medical image processing technologies, and in particular, to a method and apparatus for identifying a lesion area, an electronic device, and a storage medium. Background In recent years, the development of medical image technology is rapid, and image scanning has become a common auxiliary diagnosis and treatment means in modern medical diagnosis and treatment. After the scanning equipment is used for scanning the image of the human body, the visual medical image can be obtained through the AI auxiliary diagnosis technology, more visual physiological tissue information is provided for doctors, and the diagnosis efficiency is improved. Along with the acceleration of daily work and life rhythm, the pressure born by females gradually increases, and the incidence rate of female common diseases such as breast nodules, breast cancers, ovarian cysts, uterine tumors and the like also increases. Breast cancer is a disease in which abnormal breast cell growth is uncontrolled and forms tumors. As cancer cells grow larger, their shape becomes more complex. However, the boundary of cancer cells in medical images is often blurred, and in the process of dividing and identifying physiological tissues by scanning imaging, a cancer region is easily mistakenly identified as a plurality of cancer regions. Wrong cancerous region references can lead to inaccurate judgment of the cancerous degree by a doctor in diagnosis, and the risk of misdiagnosis exists. Disclosure of Invention The invention aims to provide a lesion area identification method, a lesion area identification device, electronic equipment and a storage medium, which can at least improve the identification degree of a lesion boundary, are favorable for ensuring the accuracy of benign and malignant judgment of a lesion and reduce the risk of misdiagnosis. In order to solve the above technical problem, in a first aspect, the present invention provides a lesion area identifying method, which at least includes: Acquiring a medical image of any human body component, and dividing and identifying lesions in the medical image to obtain a plurality of lesion areas and at least one transition area; selecting any lesion area as a target lesion area, and taking each lesion area adjacent to the target lesion area as a pending lesion area to determine a reference distance between the target lesion area and each pending lesion area; Combining the target lesion areas with any transition area to obtain combined lesion areas, determining the actual distance between each combined lesion area and the corresponding undetermined lesion area, and further determining whether the corresponding target lesion areas, transition areas and undetermined lesion areas belong to the same real lesion area or not based on the reference distance and the actual distance. Optionally, after the combining the target lesion area and any one of the transition areas to obtain a merged lesion area, determining an actual distance between each merged lesion area and the corresponding pending lesion area, and further determining whether the corresponding target lesion area, the transition area, and the pending lesion area belong to the same real lesion area based on the reference distance and the actual distance, at least further including: when the corresponding target lesion area, the transition area and the undetermined lesion area belong to the same real lesion area, determining at least lesion volume and uniformity according to the real lesion area; and acquiring at least one lesion sign index except the lesion volume and the uniformity, and judging the degree of deterioration of the lesion according to at least one lesion sign index, the lesion volume and the uniformity. Optionally, the segmentation recognition of lesions in the medical image is achieved at least by a preset lesion segmentation model configured with a confidence value; The confidence value corresponding to the background area in the medical image is at least a first threshold value; the confidence value corresponding to the lesion area in the medical image is at least not lower than a second threshold value; The confidence value corresponding to the transition region in the medical image is at least between a third threshold and a fourth threshold. Optionally, the first threshold is smaller than the third threshold, the third threshold is smaller than the fourth threshold, and the fourth threshold is not larger than the second threshold. Optionally, the reference distance refers to at least the shortest distance between the target lesion area and each of the pending lesions areas. Optionally, when the actual distance is smaller than the reference distance, determining that the corresponding target lesion region, the transition region and the pending lesion region belong to the same real lesion r