Non-local Neural Networks (2017-11-21T00:00:00.000000Z)

TL;DR

This paper presents non-local operations as a generic family of building blocks for capturing long-range dependencies in computer vision and improves object detection/segmentation and pose estimation on the COCO suite of tasks.

Abstract

Both convolutional and recurrent operations are building blocks that process one local neighborhood at a time. In this paper, we present non-local operations as a generic family of building blocks for capturing long-range dependencies. Inspired by the classical non-local means method [4] in computer vision, our non-local operation computes the response at a position as a weighted sum of the features at all positions. This building block can be plugged into many computer vision architectures. On the task of video classification, even without any bells and whistles, our nonlocal models can compete or outperform current competition winners on both Kinetics and Charades datasets. In static image recognition, our non-local models improve object detection/segmentation and pose estimation on the COCO suite of tasks. Code will be made available.

Authors

A. Gupta

21 papers

Kaiming He

22 papers

Ross B. Girshick

23 papers

TL;DR

Abstract

Authors

References57 items

Learning Affinity via Spatial Propagation Networks

Dense and Low-Rank Gaussian CRFs Using Deep Embeddings

Segmentation-Aware Convolutional Networks Using Local Attention Masks

Revisiting the Effectiveness of Off-the-shelf Temporal Modeling Approaches for Large-scale Video Classification

Attention is All you Need

Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour

A simple neural network module for relational reasoning

Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset

The Kinetics Human Action Video Dataset

Convolutional Sequence to Sequence Learning

Mask R-CNN

Asynchronous Temporal Fields for Action Recognition

Feature Pyramid Networks for Object Detection

Interaction Networks for Learning about Objects, Relations and Physics

Non-local Color Image Denoising with Convolutional Neural Networks

Aggregated Residual Transformations for Deep Neural Networks

Spatiotemporal Residual Networks for Video Action Recognition

The microsoft 2016 conversational speech recognition system

WaveNet: A Generative Model for Raw Audio

Hollywood in Homes: Crowdsourcing Data Collection for Activity Understanding

Deep Residual Learning for Image Recognition

Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks

Action recognition with trajectory-pooled deep-convolutional descriptors

Beyond short snippets: Deep networks for video classification

Fully Connected Deep Structured Networks

Conditional Random Fields as Recurrent Neural Networks

Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift

Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification

Semantic Image Segmentation with Deep Convolutional Nets and Fully Connected CRFs

Learning Spatiotemporal Features with 3D Convolutional Networks

Long-term recurrent convolutional networks for visual recognition and description

Very Deep Convolutional Networks for Large-Scale Image Recognition

ImageNet Large Scale Visual Recognition Challenge

Two-Stream Convolutional Networks for Action Recognition in Videos

Microsoft COCO: Common Objects in Context

Image denoising with multi-layer perceptrons, part 2: training trade-offs and analysis of their mechanisms

Improving neural networks by preventing co-adaptation of feature detectors

Image denoising: Can plain neural networks compete with BM3D?

Efficient Inference in Fully Connected CRFs with Gaussian Edge Potentials

Rectified Linear Units Improve Restricted Boltzmann Machines

Super-resolution from a single image

PatchMatch: a randomized correspondence algorithm for structural image editing

Image Denoising by Sparse 3-D Transform-Domain Collaborative Filtering

A non-local algorithm for image denoising

Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data

Texture synthesis by non-parametric sampling

Bilateral filtering for gray and color images

Long Short-Term Memory

Backpropagation Applied to Handwritten Zip Code Recognition

Learning representations by back-propagating errors

Multi-agent Predictive Modeling with Attentional CommNets

Learning afﬁnity via spatial propagation networks In Neural Information Processing Systems (NIPS)

Visual interaction networks

The Graph Neural Network Model

Neocognitron: A Self-Organizing Neural Network Model for a Mechanism of Visual Pattern Recognition

Author manuscript, published in "International Conference on Computer Vision (2013)" Action Recognition with Improved Trajectories

Ieee Transactions on Pattern Analysis and Machine Intelligence 1 3d Convolutional Neural Networks for Human Action Recognition

Field of Study

Journal Information

Name

Page