SimAM: A Simple, Parameter-Free Attention Module for Convolutional Neural Networks
Explore this paper's citation graph
Summary
In contrast to existing channel-wise and spatial-wise attention modules, this module instead infers 3-D attention weights for the feature map in a layer without adding parameters to the original networks.
- Type
- article
- Published
- 2021-07-18
- Cited by
- 1,849
- References
- 66
- OpenAlex
- https://openalex.org/W3166716987
- Semantic Scholar
- https://api.semanticscholar.org/CorpusID:235825945
Keywords
Simple (philosophy), Computer science, Convolutional neural network, Artificial intelligence
References
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
- Understanding the difficulty of training deep feedforward neural networks
- Very Deep Convolutional Networks for Large-Scale Image Recognition
- Painting-to-3D model alignment via discriminative visual elements
- Return of the Devil in the Details: Delving Deep into Convolutional Nets
- Sensory gain control (amplification) as a mechanism of selective attention: electrophysiological and neuroimaging evidence.
- Early and Late Mechanisms of Surround Suppression in Striate Cortex of Macaque
- Dropout: a simple way to prevent neural networks from overfitting
- Going deeper with convolutions
- Learning long-term dependencies with gradient descent is difficult
- Gradient-based learning applied to document recognition
- Visual attention: The past 25 years
- ImageNet Large Scale Visual Recognition Challenge
- Attentional modulation of visual processing.
- A taxonomy of external and internal attention.
- ImageNet classification with deep convolutional neural networks
- Deep Residual Learning for Image Recognition
- Convolutional Two-Stream Network Fusion for Video Action Recognition
- Xception: Deep Learning with Depthwise Separable Convolutions
- Aggregated Residual Transformations for Deep Neural Networks
Cited by
- FP-Age: Leveraging Face Parsing Attention for Facial Age Estimation in the Wild
- Simple Attention Module Based Speaker Verification with Iterative Noisy Label Detection
- Attention mechanisms in computer vision: A survey
- Melanoma Detection based on Swin Transformer and SimAM
- An Efficient Extreme-Exposure Image Fusion Method
- Fast writer adaptation with style extractor network for handwritten text recognition
- Lightweight Multi-Scale Asymmetric Attention Network for Image Super-Resolution
- Contrastive Learning from Extremely Augmented Skeleton Sequences for Self-supervised Action Recognition
- Wavelet-Attention CNN for image classification
- Feature Attention Parallel Aggregation Network for Single Image Haze Removal
- IL-MCAM: An interactive learning and multi-channel attention mechanism-based weakly supervised colorectal histopathology image classification approach
- Overview of One-Dimensional Continuous Functions with Fractional Integral and Applications in Reinforcement Learning
- ReplaceBlock: An improved regularization method based on background information
- The Fixed Sub-Center: A Better Way to Capture Data Complexity
- Multi-scale spatial-spectral attention network for multispectral image compression based on variational autoencoder
- Attentive Manifold Mixup for Model Robustness
- Object Detection by Attention-Guided Feature Fusion Network
- Deep parameter-free attention hashing for image retrieval
- CitrusYOLO: A Algorithm for Citrus Detection under Orchard Environment Based on YOLOv4
- MSHT: Multi-Stage Hybrid Transformer for the ROSE Image Analysis of Pancreatic Cancer
Related papers
- A study on the application of convolutional neural network for the acceleration of nanoscale device simulation
- On-Chip Optical Convolutional Neural Networks
- Increasing of convolutional neural network performance using residue number system
- Uniform Experimental Design for Optimizing the Parameters of Multi-input Convolutional Neural Networks
- Accurate and compact convolutional neural network based on stochastic computing
- Study of Quantized Hardware Deep Neural Networks Based on Resistive Switching Devices, Conventional versus Convolutional Approaches