Food image segmentation method and system based on dynamic transformer

Abstract

The invention relates to a food image segmentation method and system based on dynamic transformer, and the method comprises the steps: S1, dividing an input food image into a series of image blocks with different sizes, and inputting the image blocks into a plurality of dynamic visual transformer encoder networks with different sizes; outputting multi-layer image feature vectors of different scales; s2, extracting the image feature vectors of the preset layer for fusion to obtain fused image feature vectors; s3, constructing a multi-stage feature aggregation network, performing top-down feature fusion on the fused image feature vectors, constructing a feature pyramid, and obtaining multi-scale feature fusion vectors; and S4, constructing a segmentation decoder, performing convolution and up-sampling operation on the multi-scale features fused by the feature pyramid, and finally generating a segmentation result with accurate food category boundary segmentation. The method provided by the invention can be adaptive to different picture scales, and richness and integrity of picture semantic information extraction are improved, so that a food segmentation model has generalization and robustness.

Description

technical field [0001] The invention relates to the fields of computer vision and food computing, in particular to a method and system for food image segmentation based on a dynamic transformer. Background technique [0002] Computer vision is emerging as an emerging technology for acquiring and analyzing images of real scenes, which helps intelligent systems perceive the world from images and multidimensional data. The core technology of computer vision has always been related to image analysis and processing, which can classify, detect and segment some specific objects in images. Image semantic segmentation makes pixel-level predictions on a set of object categories; segmentation is often a more demanding task than image classification, which predicts labels for an entire image. From the earliest traditional methods, such as thresholding, k-means clustering, and region growing, to some deep learning models that have achieved good results, such as FCN, PSPNet, and DeepLab ...

AI Technical Summary

Problems solved by technology

On the other hand, the distribution of food datasets is usually unbalanced, and there is a long-tail distribution problem

Images