TasksSotADatasetsPapersMethodsSubmitAbout
Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable BenchmarksAll SotADatasetsPapersMethods

Community

Submit ResultsAbout

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Methods/Res2Net Block

Res2Net Block

Computer VisionIntroduced 200025 papers
Source Paper

Description

A Res2Net Block is an image model block that constructs hierarchical residual-like connections within one single residual block. It was proposed as part of the Res2Net CNN architecture.

The block represents multi-scale features at a granular level and increases the range of receptive fields for each network layer. The 3×33 \times 33×3 filters of nnn channels is replaced with a set of smaller filter groups, each with www channels. These smaller filter groups are connected in a hierarchical residual-like style to increase the number of scales that the output features can represent. Specifically, we divide input feature maps into several groups. A group of filters first extracts features from a group of input feature maps. Output features of the previous group are then sent to the next group of filters along with another group of input feature maps.

This process repeats several times until all input feature maps are processed. Finally, feature maps from all groups are concatenated and sent to another group of 1×11 \times 11×1 filters to fuse information altogether. Along with any possible path in which input features are transformed to output features, the equivalent receptive field increases whenever it passes a 3×33 \times 33×3 filter, resulting in many equivalent feature scales due to combination effects.

One way of thinking of these blocks is that they expose a new dimension, scale, alongside the existing dimensions of depth, width, and cardinality.

Papers Using This Method

Nes2Net: A Lightweight Nested Architecture for Foundation Model Driven Speech Anti-spoofing2025-04-08ERes2NetV2: Boosting Short-Duration Speaker Verification Performance with Computational Efficiency2024-06-04Speech enhancement deep-learning architecture for efficient edge processing2024-05-27Enhancing Retinal Vascular Structure Segmentation in Images With a Novel Design Two-Path Interactive Fusion Module Model2024-03-03NeXt-TDNN: Modernizing Multi-Scale Temporal Convolution Backbone for Speaker Verification2023-12-14Improving Short Utterance Anti-Spoofing with AASIST22023-09-15A region and category confidence-based multi-task network for carotid ultrasound image segmentation and classification2023-07-02Multi-perspective Information Fusion Res2Net with RandomSpecmix for Fake Speech Detection2023-06-27An Enhanced Res2Net with Local and Global Feature Fusion for Speaker Verification2023-05-22Breaking Through the Haze: An Advanced Non-Homogeneous Dehazing Method based on Fast Fourier Convolution and ConvNeXt2023-05-08Cross-modal information fusion for voice spoofing detection2023-02-01Synthetic Voice Detection and Audio Splicing Detection using SE-Res2Net-Conformer Architecture2022-10-07Breast Cancer Classification Based on Histopathological Images Using a Deep Learning Capsule Network2022-08-01Frequency and Multi-Scale Selective Kernel Attention for Speaker Verification2022-04-03MFSNet: A Multi Focus Segmentation Network for Skin Lesion Segmentation2022-03-27Phase-Aware Spoof Speech Detection Based on Res2Net with Phase Network2022-03-21Pushing the limits of raw waveform speaker recognition2022-03-16Channel-wise Gated Res2Net: Towards Robust Detection of Synthetic Speech Attacks2021-07-19DW-GAN: A Discrete Wavelet Transform GAN for NonHomogeneous Dehazing2021-04-18Partially-Connected Differentiable Architecture Search for Deepfake and Spoofing Detection2021-04-07