Detectors++: The Robust Baseline for a Defect Detection System

Fityanul Akhyar, Chih Yang Lin, Gugan S. Kathiresan, Bharath Surianarayanan, Chao Yung Hsu

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

3 Scopus citations

Abstract

Focusing on the task of steel surface defect localization, this study employs the latest state-of-the-art RCNN family, Cascade RCNN, on top of the current FPN model called DetectoRS-ResNeXt. The baseline was tested individually using Side Aware Boundary Localization (SABL) plus pixel domain augmentation to obtain the precision of predictions. Trained on a well-known real-world dataset, Severstal, our proposal achieves a mAP of 82.5% which offers the potential to serve as a high-quality defect detection baseline.

Original languageEnglish
Title of host publication2021 IEEE International Conference on Consumer Electronics-Taiwan, ICCE-TW 2021
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781665433280
DOIs
StatePublished - 2021
Event8th IEEE International Conference on Consumer Electronics-Taiwan, ICCE-TW 2021 - Penghu, Taiwan
Duration: 15 Sep 202117 Sep 2021

Publication series

Name2021 IEEE International Conference on Consumer Electronics-Taiwan, ICCE-TW 2021

Conference

Conference8th IEEE International Conference on Consumer Electronics-Taiwan, ICCE-TW 2021
Country/TerritoryTaiwan
CityPenghu
Period15/09/2117/09/21

Fingerprint

Dive into the research topics of 'Detectors++: The Robust Baseline for a Defect Detection System'. Together they form a unique fingerprint.

Cite this