HarDNet: A Low Memory Traffic Network

@article{Chao2019HarDNetAL,
  title={HarDNet: A Low Memory Traffic Network},
  author={Ping Chao and Chao-Yang Kao and Yu-Shan Ruan and Chien-Hsiang Huang and Youn-Long Lin},
  journal={2019 IEEE/CVF International Conference on Computer Vision (ICCV)},
  year={2019},
  pages={3551-3560}
}
  • Ping Chao, Chao-Yang Kao, +2 authors Youn-Long Lin
  • Published in
    IEEE/CVF International…
    2019
  • Computer Science
  • State-of-the-art neural network architectures such as ResNet, MobileNet, and DenseNet have achieved outstanding accuracy over low MACs and small model size counterparts. However, these metrics might not be accurate for predicting the inference time. We suggest that memory traffic for accessing intermediate feature maps can be a factor dominating the inference latency, especially in such tasks as real-time object detection and semantic segmentation of high-resolution video. We propose a Harmonic… CONTINUE READING

    Figures, Tables, and Topics from this paper.

    Citations

    Publications citing this paper.
    SHOWING 1-10 OF 10 CITATIONS

    RGPNet: A Real-Time General Purpose Semantic Segmentation

    VIEW 5 EXCERPTS
    CITES METHODS & BACKGROUND
    HIGHLY INFLUENCED

    Distilled Semantics for Comprehensive Scene Understanding from Videos

    VIEW 5 EXCERPTS
    CITES BACKGROUND & METHODS
    HIGHLY INFLUENCED

    CSPNet: A New Backbone that can Enhance Learning Capability of CNN

    VIEW 6 EXCERPTS
    CITES METHODS
    HIGHLY INFLUENCED

    Adapting JPEG XS gains and priorities to tasks and contents

    VIEW 1 EXCERPT
    CITES METHODS

    Real-Time 3D Door Detection and Classification on a Low-Power Device

    VIEW 3 EXCERPTS

    Segmenting Transparent Objects in the Wild

    VIEW 1 EXCERPT
    CITES METHODS

    YOLOv4: Optimal Speed and Accuracy of Object Detection

    VIEW 1 EXCERPT
    CITES BACKGROUND

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 37 REFERENCES

    Sparsely Aggregated Convolutional Networks

    VIEW 6 EXCERPTS
    HIGHLY INFLUENTIAL

    Densely Connected Convolutional Networks

    VIEW 5 EXCERPTS
    HIGHLY INFLUENTIAL

    Semantic object classes in video: A high-definition ground truth database

    VIEW 2 EXCERPTS
    HIGHLY INFLUENTIAL

    NullHop: A Flexible Convolutional Neural Network Accelerator Based on Sparse Representations of Feature Maps

    VIEW 2 EXCERPTS

    MobileNetV2: Inverted Residuals and Linear Bottlenecks

    VIEW 2 EXCERPTS

    Single-Shot Refinement Neural Network for Object Detection