Cell image segmentation model

Abstract

The invention discloses a cell image segmentation model, which takes a U-Net network as a basic framework and comprises an encoder, a decoder and an attention mechanism. A residual block is adopted toreplace the encoder of the U-Net network and a convolutional layer in the decoder.

Description

technical field [0001] The invention relates to the field of aerospace technology, in particular to a cell image segmentation model. Background technique [0002] Cells are the basic unit of the structure and function of organisms. In-depth research on cells is the key to uncovering the mysteries of life, transforming life and conquering diseases. With the development of technology, computer-aided technology has been widely used in the automatic analysis of cell images. In cell image analysis, cell segmentation, as one of the most basic and important fields, is the basic premise for identifying, counting, and measuring cell images. [0003] Under the influence of some objective reasons, the current optical imaging system has the problems of uneven brightness and low contrast between cells and background in the collected phase-contrast microscopic cell images, which greatly increases the difficulty of segmentation. In response to the above problems, Yin et al. analyzed the ...

AI Technical Summary

Problems solved by technology

For example, Yi et al. combined SSD and U-Net network for instance segmentation of neural cells, and adopted an attention mechanism in the detection and segmentation modules to focus the model on task-related features; and Alom proposed a recursive residual neural network R2U-Net based on the U-Net network, the residual unit can slow down the gradient disappearance when training the deep network, and the accumulation of recursive residual features can ensure that the network achieves better segmentation tasks. Feature representation, but using this method for image segmentation, weak edge segmentation often occurs

Images