Cover Image

Fine-tuned Surface Object Detection Applying Pre-trained Mask R-CNN Models

This study evaluates road surface object detection tasks using four Mask R-CNN models available on the Tensor-Flow Object Detection API. The models were pre-trained using COCO datasets and fine-tuned by 15,1SS segmented road surface annotation tags. Validation data set was used to obtain Average Pre...

Full description

Saved in:
Bibliographic Details
Main Authors: Fujita, H., Itagaki, M., Ichikawa, K., Hooi, Y.K., Kawahara, K., Sarlan, A.
Format: Conference or Workshop Item
Published: Institute of Electrical and Electronics Engineers Inc. 2020
Online Access:https://www.scopus.com/inward/record.uri?eid=2-s2.0-85097532534&doi=10.1109%2fICCI51257.2020.9247666&partnerID=40&md5=2758293eafa9dbd7d5a0e244112984ec
http://eprints.utp.edu.my/29862/
Tags: Add Tag
No Tags, Be the first to tag this record!
id my.utp.eprints.29862
record_format eprints
spelling my.utp.eprints.298622022-03-25T03:04:45Z Fine-tuned Surface Object Detection Applying Pre-trained Mask R-CNN Models Fujita, H. Itagaki, M. Ichikawa, K. Hooi, Y.K. Kawahara, K. Sarlan, A. This study evaluates road surface object detection tasks using four Mask R-CNN models available on the Tensor-Flow Object Detection API. The models were pre-trained using COCO datasets and fine-tuned by 15,1SS segmented road surface annotation tags. Validation data set was used to obtain Average Precisions and Average Recalls. Result indicates a substantial false negatives or "left judgement"counts for linear cracks, joints, fillings, potholes, stains, shadows and patching with grid cracks classes. There were significant number of incorrectly predicted label instances. To improve the result, an alternative metric calculation method was tested. However, the results showed strong mutual interferences caused by misinterpretation of the scratches with other object classes. © 2020 IEEE. Institute of Electrical and Electronics Engineers Inc. 2020 Conference or Workshop Item NonPeerReviewed https://www.scopus.com/inward/record.uri?eid=2-s2.0-85097532534&doi=10.1109%2fICCI51257.2020.9247666&partnerID=40&md5=2758293eafa9dbd7d5a0e244112984ec Fujita, H. and Itagaki, M. and Ichikawa, K. and Hooi, Y.K. and Kawahara, K. and Sarlan, A. (2020) Fine-tuned Surface Object Detection Applying Pre-trained Mask R-CNN Models. In: UNSPECIFIED. http://eprints.utp.edu.my/29862/
institution Universiti Teknologi Petronas
building UTP Resource Centre
collection Institutional Repository
continent Asia
country Malaysia
content_provider Universiti Teknologi Petronas
content_source UTP Institutional Repository
url_provider http://eprints.utp.edu.my/
description This study evaluates road surface object detection tasks using four Mask R-CNN models available on the Tensor-Flow Object Detection API. The models were pre-trained using COCO datasets and fine-tuned by 15,1SS segmented road surface annotation tags. Validation data set was used to obtain Average Precisions and Average Recalls. Result indicates a substantial false negatives or "left judgement"counts for linear cracks, joints, fillings, potholes, stains, shadows and patching with grid cracks classes. There were significant number of incorrectly predicted label instances. To improve the result, an alternative metric calculation method was tested. However, the results showed strong mutual interferences caused by misinterpretation of the scratches with other object classes. © 2020 IEEE.
format Conference or Workshop Item
author Fujita, H.
Itagaki, M.
Ichikawa, K.
Hooi, Y.K.
Kawahara, K.
Sarlan, A.
spellingShingle Fujita, H.
Itagaki, M.
Ichikawa, K.
Hooi, Y.K.
Kawahara, K.
Sarlan, A.
Fine-tuned Surface Object Detection Applying Pre-trained Mask R-CNN Models
author_facet Fujita, H.
Itagaki, M.
Ichikawa, K.
Hooi, Y.K.
Kawahara, K.
Sarlan, A.
author_sort Fujita, H.
title Fine-tuned Surface Object Detection Applying Pre-trained Mask R-CNN Models
title_short Fine-tuned Surface Object Detection Applying Pre-trained Mask R-CNN Models
title_full Fine-tuned Surface Object Detection Applying Pre-trained Mask R-CNN Models
title_fullStr Fine-tuned Surface Object Detection Applying Pre-trained Mask R-CNN Models
title_full_unstemmed Fine-tuned Surface Object Detection Applying Pre-trained Mask R-CNN Models
title_sort fine-tuned surface object detection applying pre-trained mask r-cnn models
publisher Institute of Electrical and Electronics Engineers Inc.
publishDate 2020
url https://www.scopus.com/inward/record.uri?eid=2-s2.0-85097532534&doi=10.1109%2fICCI51257.2020.9247666&partnerID=40&md5=2758293eafa9dbd7d5a0e244112984ec
http://eprints.utp.edu.my/29862/
_version_ 1738657025972240384
score 13.211869

Similar Items