Gaussian Filter in CRF Based Semantic Segmentation

TL;DR

This paper introduces a multi-resolution neural network for fully convolutional networks (FCN) that applies Gaussian filtering to improve the accuracy and speed of CRF-based semantic segmentation.

Contribution

It proposes a novel Gaussian filtering approach in CRF kernels to enhance segmentation precision and training efficiency in neural network models.

Findings

Improved segmentation accuracy with Gaussian filtering.

Faster training times for CRF-based models.

Reduced oscillations in segmentation results.

Abstract

Artificial intelligence is making great changes in academy and industry with the fast development of deep learning, which is a branch of machine learning and statistical learning. Fully convolutional network [1] is the standard model for semantic segmentation. Conditional random fields coded as CNN [2] or RNN [3] and connected with FCN has been successfully applied in object detection [4]. In this paper, we introduce a multi-resolution neural network for FCN and apply Gaussian filter to the extended CRF kernel neighborhood and the label image to reduce the oscillating effect of CRF neural network segmentation, thus achieve higher precision and faster training speed.