一元运算
计算机科学
帕斯卡(单位)
人工智能
块(置换群论)
分割
成对比较
期限(时间)
编码(集合论)
卷积神经网络
背景(考古学)
像素
人工神经网络
对象(语法)
模式识别(心理学)
数学
古生物学
集合(抽象数据类型)
物理
程序设计语言
组合数学
生物
量子力学
几何学
作者
Minghao Yin,Zhuliang Yao,Yue Cao,Xiu Li,Zheng Zhang,Stephen Lin,Han Hu
标识
DOI:10.1007/978-3-030-58555-6_12
摘要
The non-local block is a popular module for strengthening the context modeling ability of a regular convolutional neural network. This paper first studies the non-local block in depth, where we find that its attention computation can be split into two terms, a whitened pairwise term accounting for the relationship between two pixels and a unary term representing the saliency of every pixel. We also observe that the two terms trained alone tend to model different visual clues, e.g. the whitened pairwise term learns within-region relationships while the unary term learns salient boundaries. However, the two terms are tightly coupled in the non-local block, which hinders the learning of each. Based on these findings, we present the disentangled non-local block, where the two terms are decoupled to facilitate learning for both terms. We demonstrate the effectiveness of the decoupled design on various tasks, such as semantic segmentation on Cityscapes, ADE20K and PASCAL Context, object detection on COCO, and action recognition on Kinetics. Code is available at https://github.com/yinmh17/DNL-Semantic-Segmentation and https://github.com/Howal/DNL-Object-Detection
科研通智能强力驱动
Strongly Powered by AbleSci AI