A Weakly Supervised Semantic Segmentation Method and System Based on Graffiti

Abstract

The present invention proposes a weakly supervised semantic segmentation method and system based on graffiti, including: obtaining multiple training pictures, each of which has corresponding graffiti marks and edge maps; selecting the training picture as the current picture, and inputting the current picture to the semantic segmentation network , to obtain the high-level semantic features of the current picture; input the high-level semantic features to the prediction and correction network to obtain the segmentation result map of the current picture, and according to the graffiti marks of the current picture, obtain the cross-entropy loss of the graffiti mark area in the current picture; the high-level semantic Input the feature to the boundary regression network to obtain the boundary map of the target in the current picture, and according to the edge map of the current picture, obtain the mean variance loss of the boundary area in the boundary map; construct a total loss function, and judge whether the total loss function converges, if so, Then, the current prediction and correction network is used as the semantic segmentation model; the image to be semantically segmented is input to the semantic segmentation model, and the segmentation result map of the image to be semantically segmented is obtained.

Description

technical field [0001] The method belongs to the field of machine learning and computer vision, and in particular relates to machine learning problems for weakly supervised semantic segmentation in computer vision. Background technique [0002] The current popular scene segmentation methods are mainly based on the Fully Convolutional Network (FCN) and its variants. These methods all combine the idea of ​​transfer learning, using the pre-trained convolutional neural network on a large-scale image classification dataset, adjusting it to a fully convolutional network structure and retraining on a weakly supervised semantic segmentation dataset. For fine-label training, this method can achieve good segmentation results. However, when only weakly labeled training networks are provided, such methods mainly suffer from the following problems: (1) Inconsistencies and discontinuities often appear in the segmentation results, and (2) the segmentation boundaries of objects are often i...

AI Technical Summary

Problems solved by technology

However, when only weakly labeled training networks are provided, such methods mainly suffer from the following problems: (1) Inconsistencies and discontinuities often appear in the segmentation results, and (2) the segmentation boundaries of objects are often imprecise and incoherent

The segmentation model trained directly with graffiti marks can only produce rough segmentation results, mainly because the graffiti marks only contain part of the semantic information and do not provide fine boundary information to guide the model to accurately segment each target

Images