ResNeSt: Split-Attention Networks

Abstract

It is well known that featuremap attention and multi-path representation are important for visual recognition. In this paper, we present a modularized architecture, which applies the channel-wise attention on different network branches to leverage their success in capturing cross-feature interactions and learning diverse representations. Our design results in a simple and unified computation block, which can be parameterized using only a few variables. Our model, named ResNeSt, outperforms EfficientNet in accuracy and latency trade-off on image classification. In addition, ResNeSt has achieved superior transfer learning results on several public benchmarks serving as the backbone, and has been adopted by the winning entries of COCO-LVIS challenge. The source code for complete system and pretrained models are publicly available.

Cited by in corpus (26)

• Automated Pavement Crack Segmentation Using U-Net-based Convolutional Neural Network
• Compute and Energy Consumption Trends in Deep Learning Inference
• Hardware and Software Optimizations for Accelerating Deep Neural Networks: Survey of Current Trends, Challenges, and the Road Ahead
• CBNet: A Composite Backbone Network Architecture for Object Detection
• Causal Contextual Prediction for Learned Image Compression
• Communication optimization strategies for distributed deep neural network training: A survey
• Attentional Feature Refinement and Alignment Network for Aircraft Detection in SAR Imagery
• Towards Better Accuracy-efficiency Trade-offs: Divide and Co-training
• Graph CNN for Moving Object Detection in Complex Environments from Unseen Videos
• Contrastive Multiview Coding with Electro-optics for SAR Semantic Segmentation
• A Novel Upsampling and Context Convolution for Image Semantic Segmentation
• CVEGAN: A Perceptually-inspired GAN for Compressed Video Enhancement
• Boosting EfficientNets Ensemble Performance via Pseudo-Labels and Synthetic Images by pix2pixHD for Infection and Ischaemia Classification in Diabetic Foot Ulcers
• Video Relation Detection with Trajectory-aware Multi-modal Features
• Towards Counterfactual and Contrastive Explainability and Transparency of DCNN Image Classifiers
• OCTAve: 2D en face Optical Coherence Tomography Angiography Vessel Segmentation in Weakly-Supervised Learning with Locality Augmentation
• FedDCT: Federated Learning of Large Convolutional Neural Networks on Resource Constrained Devices using Divide and Collaborative Training
• Fusion Flow-enhanced Graph Pooling Residual Networks for Unmanned Aerial Vehicles Surveillance in Day and Night Dual Visions
• Distilling Ensemble of Explanations for Weakly-Supervised Pre-Training of Image Segmentation Models
• All Attention U-NET for Semantic Segmentation of Intracranial Hemorrhages In Head CT Images
• Learning from the few: Fine-grained approach to pediatric wrist pathology recognition on a limited dataset
• An Effective Deep Network for Head Pose Estimation without Keypoints
• Navigating limitations with precision: A fine-grained ensemble approach to wrist pathology recognition on a limited x-ray dataset
• Multi-view Feature Augmentation with Adaptive Class Activation Mapping
• Feature boosting with efficient attention for scene parsing
• UNet--: Memory-Efficient and Feature-Enhanced Network Architecture based on U-Net with Reduced Skip-Connections