Kidney tumor segmentation method based on multi-scale feature learning

Abstract

The invention discloses a kidney tumor segmentation method based on multi-scale feature learning, and the method comprises the following steps: S1, obtaining an abdomen scanning image, and randomly dividing the obtained abdomen scanning image into a training set; S2, preprocessing the abdomen scanning images in the training set to obtain preprocessed images; S3, constructing a multi-scale featurenetwork, and fully capturing global structure information of the image through the network in combination with a pyramid pooling module and a feature pyramid fusion module so as to perform accurate kidney segmentation; S4, predicting and segmenting the image preprocessed in the step S2 through a multi-scale feature network; The invention can be used for effectively detecting kidney tumors with different volumes, false negative results are avoided, and high-accuracy detection results are obtained.

Description

technical field [0001] The present invention relates to the technical field of medical image processing, in particular to a renal tumor segmentation method based on multi-scale feature learning. Background technique [0002] The kidney is an important organ of the human body. Once the kidney function is damaged, a variety of metabolic end products will accumulate in the body, which will affect life safety. Among various kidney diseases, kidney tumor is the number one dangerous disease for kidney health. At present, CT imaging examination is one of the main examination methods for renal tumors and other renal diseases. According to the size of renal tumors, doctors can grade the severity of tumors and formulate corresponding treatment methods; at the same time, they can locate renal tumors and analyze their shape and size; the existing accurate segmentation and judgment of the kidney and renal tumor area on the acquired kidney image through medical image processing has effec...

AI Technical Summary

Problems solved by technology

At present, CT imaging examination is one of the main examination methods for renal tumors and other renal diseases. According to the size of renal tumors, doctors can grade the severity of tumors and formulate corresponding treatment methods; at the same time, they can locate renal tumors and analyze their shape and size; the existing accurate segmentation and judgment of the kidney and renal tumor area on the acquired kidney image through medical image processing has effectively alleviated the workload of doctors and demonstrated the effectiveness of technology intelligence. The existing segmentation methods commonly used detection segmentation The network is a U-Net network, but there is a large heterogeneity among renal tumors, which is manifested by large differences in the shape and size of different renal tumors, which makes it difficult for the U-Net network to robustly learn the characteristics of each tumor. For small-volume renal tumors, it is easy to produce false negative results, and it is difficult to obtain high-accuracy test results

Images