Shulou(Shulou.com)06/02 Report--

Written by Amusi

Date：2019-11-28

Weixin Official Accounts: CVer

Link: Big Inventory| 4 Best Object Detection Algorithms for 2019

Last time, I compiled a summary of recent outstanding papers on target detection. For details, see: Review 8 latest papers on target detection (EfficientDet/EdgeNet/ASFF/RoIMix, etc.). A lot of CVers in the background and WeChat community reflect: These are very new papers, I just started can not understand how to do?

A: Look at the summary! After roughly sorting out the context, pick up the paper and read the code carefully.

It's just the end of November 2019, and it's time to do a summative inventory, which is the 2019 Object Detection Review. If CVer likes this kind of inventory, please give this article a like, if there are many people like it, other CV directions (segmentation/tracking, etc.) will also be launched as soon as possible!

The target detection papers shared in this article will be synchronously pushed to github. Welcome to star/fork (click to read the original text, or visit it directly):

amusi/awesome-object-detection github.com

Target detection papers

【1】Object Detection in 20 Years: A Survey

Date: May 2019

Authors: University of Michigan & Beihang University & Carlton University & Didi Chuxing

Link: https://arxiv.org/abs/1905.0505 5

Recommended index: ★★ ★★★

Note: 39 pages of target detection review, 411 references in total, too strong!

Target Detection Milestones: 2001-2019 Target Detection Multiscale Methods: 2001 - 2019 Target Detection Bounding Regression Methods: 2001 - 2019 Target Detection Non-Maximum Suppression (NMS) Methods: 1994 - 2019

【2】A Survey of Deep Learning-based Object Detection

Date: July 2019

Author: Xidian University

Link: https://arxiv.org/abs/1907.0940 8

Recommended index: ★★★★

Note: The 30-page target detection review, from Fast R-CNN to NAS-FPN, presents mAP data on COCO datasets, introducing more than 10 datasets, totaling 317 references!

Two-stage and One-stage target detection frameworks Comparison of four target detection algorithms using different size feature maps MS COCO dataset

【3】Recent Advances in Deep Learning for Object Detection

Date: August 2019

By Singapore University of Management &Salesforce

Link: https://arxiv.org/abs/1908.0367 3

Recommended index: ★★★★

Note: 40 pages of target detection review, 256 references in total! From 2013 OverFeat to 2019 NAS-FPN/CenterNet/DetNAS, covering target detection mechanisms, learning strategies, and application directions. A comprehensive comparison of algorithms under VOC/COCO datasets is also presented.

Target Detection Milestones: 2012-2019 Target Detection Key Knowledge Points VOC Dataset Algorithm Performance Comparison MS COCO Dataset Algorithm Performance Comparison

【4】Imbalance Problems in Object Detection: A Review

Date: September 2019

Middle East Technical University

Link: https://arxiv.org/abs/1909.0016 9

Recommended index: ★★★★

Note: 31 pages of target detection review with a total of 166 references! The methods of feature extraction improvement, loss function and sampling method are introduced respectively.

Balance problems Two-stage, One-stage and Bottom-Up Object Detection Foundation Framework

Target Detection Generic Framework Training Flow Feature Level Imbalance Method Example

https://zhuanlan.zhihu.com/p/94090477

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.