PVANet: Lightweight Deep Neural Networks for Real-time Object Detection

Sanghoon Hong, Byungseok Roh, Kye-Hyeon Kim, Yeongjae Cheon, Minje Park
11/23/2016
cs.CV

Official code: sanghoon/pva-faster-rcnn (649 stars, last commit 2026-02-28).

Abstract

In object detection, reducing computational cost is as important as improving accuracy for most practical usages. This paper proposes a novel network structure, which is an order of magnitude lighter than other state-of-the-art networks while maintaining the accuracy. Based on the basic principle of more layers with less channels, this new deep neural network minimizes its redundancy by adopting recent innovations including C.ReLU and Inception structure. We also show that this network can be trained efficiently to achieve solid results on well-known object detection benchmarks: 84.9% and 84.2% mAP on VOC2007 and VOC2012 while the required compute is less than 10% of the recent ResNet-101.

View on arXivView PDF

Code Implementations(1)

sanghoon/pva-faster-rcnnOfficial100%

Demo code for PVANet

649233C, C++Aug 31, 20167 months agoNOASSERTION

Similar Papers

Scaled-YOLOv4: Scaling Cross Stage Partial Network

89% match1 repo2.0k2020

Fast R-CNN

88% match1 repo3.5k2015

LeYOLO, New Embedded Architecture for Object Detection

87% match1 repo2002024

PP-PicoDet: A Better Real-Time Object Detector on Mobile Devices

87% match1 repo14.3k2021

SSD: Single Shot MultiBox Detector

86% match1 repo4.8k2015

EfficientDet: Scalable and Efficient Object Detection

86% match1 repo6.5k2019

Cite this paper

@article{hong2016pvanet,
  title  = {PVANet: Lightweight Deep Neural Networks for Real-time Object Detection},
  author = {Sanghoon Hong and Byungseok Roh and Kye-Hyeon Kim and Yeongjae Cheon and Minje Park},
  year   = {2016},
  eprint = {1611.08588},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1611.08588}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1611.08588.svg)](https://codeofpaper.com/paper/1611.08588)

Discussion