Research map / Video / Action recognition
Multi-Fiber Networks for Video Recognition
Yunpeng Chen, Yannis Kalantidis, Jianshu Li, Shuicheng Yan, Jiashi Feng
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| embed multi-fiber unit into ResNet-18 | ResNet-18 · ImageNet-1k val | Top-1 accuracy: 71.4 % -> 74.3 % | Improved |
| embed multi-fiber unit into MobileNet-v2 (1.4) | MobileNet-v2 · ImageNet-1k val | Top-1 accuracy: 72.2 % -> 73.0 % | Improved |
| add the multiplexer module between fibers (ablation) | 2D MF-Net (N=16) · ImageNet-1k val | Top-1 accuracy: 70.2 % -> 74.6 % | Improved |
| add multiplexer versus making the network deeper and wider without it (ablation) | 2D MF-Net (N=16) · ImageNet-1k val | Top-1 accuracy: 71.0 % -> 74.6 % | Improved |
| increase number of fibers from 12 to 24 (ablation) | 2D MF-Net · ImageNet-1k val | Top-1 accuracy: 74.5 % -> 75.4 % | Improved |
| increase number of fibers from 16 to 24 (ablation) | 2D MF-Net · ImageNet-1k val | Top-5 accuracy: 92.0 % -> 92.5 % | Improved |
| replace R(2+1)D-RGB with MF-Net on Kinetics | MF-Net · Kinetics val (RGB only) | Top-1 accuracy: 72.0 % -> 72.8 % | Improved |
| MF-Net versus Res3D on UCF-101 | MF-Net · UCF-101 (Kinetics pretrained) | accuracy: 85.8 % -> 96.0 % | Improved |
| MF-Net versus R(2+1)D-RGB on UCF-101 | MF-Net · UCF-101 (Kinetics pretrained) | accuracy: 96.8 % -> 96.0 % | Mixed |
| MF-Net versus CoViAR on HMDB51 | MF-Net · HMDB51 (Kinetics pretrained) | accuracy: 59.1 % -> 74.6 % | Improved |
Closest papers by meaning
- AssembleNet: Searching for Multi-Stream Neural Connectivity in Video Architectures (ICLR 2020 · similarity 0.86)
- MVFNet: Multi-View Fusion Network for Efficient Video Recognition (AAAI 2021 · similarity 0.86)
- Convolutional Two-Stream Network Fusion for Video Action Recognition (CVPR 2016 · similarity 0.85)
- Sparsely-Connected Neural Networks: Towards Efficient VLSI Implementation of Deep Neural Networks (ICLR 2017 · similarity 0.85)
- Learning Spatio-Temporal Representation With Pseudo-3D Residual Networks (ICCV 2017 · similarity 0.85)
- Rethinking Spatiotemporal Feature Learning: Speed-Accuracy Trade-offs in Video Classification (ECCV 2018 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
