FCIoU: A Targeted Approach for Improving Minority Class Detection in Semantic Segmentation Systems
In this paper, we present a comparative study of modern semantic segmentation loss functions and their resultant impact when applied with state-of-the-art off-road datasets. Class imbalance, inherent in these datasets, presents a significant challenge to off-road terrain semantic segmentation system...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-11-01
|
| Series: | Machine Learning and Knowledge Extraction |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2504-4990/5/4/85 |
| _version_ | 1797380240663642112 |
|---|---|
| author | Jonathan Plangger Mohamed Atia Hicham Chaoui |
| author_facet | Jonathan Plangger Mohamed Atia Hicham Chaoui |
| author_sort | Jonathan Plangger |
| collection | DOAJ |
| description | In this paper, we present a comparative study of modern semantic segmentation loss functions and their resultant impact when applied with state-of-the-art off-road datasets. Class imbalance, inherent in these datasets, presents a significant challenge to off-road terrain semantic segmentation systems. With numerous environment classes being extremely sparse and underrepresented, model training becomes inefficient and struggles to comprehend the infrequent minority classes. As a solution to this problem, loss functions have been configured to take class imbalance into account and counteract this issue. To this end, we present a novel loss function, Focal Class-based Intersection over Union (FCIoU), which directly targets performance imbalance through the optimization of class-based Intersection over Union (IoU). The new loss function results in a general increase in class-based performance when compared to state-of-the-art targeted loss functions. |
| first_indexed | 2024-03-08T20:34:32Z |
| format | Article |
| id | doaj.art-15331273edb74146979f70e48a6dbbd5 |
| institution | Directory Open Access Journal |
| issn | 2504-4990 |
| language | English |
| last_indexed | 2024-03-08T20:34:32Z |
| publishDate | 2023-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Machine Learning and Knowledge Extraction |
| spelling | doaj.art-15331273edb74146979f70e48a6dbbd52023-12-22T14:22:12ZengMDPI AGMachine Learning and Knowledge Extraction2504-49902023-11-01541746175910.3390/make5040085FCIoU: A Targeted Approach for Improving Minority Class Detection in Semantic Segmentation SystemsJonathan Plangger0Mohamed Atia1Hicham Chaoui2Department of Electronics (DOE), Carleton University, Ottawa, ON K1S 5B6, CanadaDepartment of Systems and Communication, Carleton University, Ottawa, ON K1S 5B6, CanadaDepartment of Electronics (DOE), Carleton University, Ottawa, ON K1S 5B6, CanadaIn this paper, we present a comparative study of modern semantic segmentation loss functions and their resultant impact when applied with state-of-the-art off-road datasets. Class imbalance, inherent in these datasets, presents a significant challenge to off-road terrain semantic segmentation systems. With numerous environment classes being extremely sparse and underrepresented, model training becomes inefficient and struggles to comprehend the infrequent minority classes. As a solution to this problem, loss functions have been configured to take class imbalance into account and counteract this issue. To this end, we present a novel loss function, Focal Class-based Intersection over Union (FCIoU), which directly targets performance imbalance through the optimization of class-based Intersection over Union (IoU). The new loss function results in a general increase in class-based performance when compared to state-of-the-art targeted loss functions.https://www.mdpi.com/2504-4990/5/4/85loss functionoff-roadsemantic segmentationclass imbalanceterrain segmentationU-Net |
| spellingShingle | Jonathan Plangger Mohamed Atia Hicham Chaoui FCIoU: A Targeted Approach for Improving Minority Class Detection in Semantic Segmentation Systems Machine Learning and Knowledge Extraction loss function off-road semantic segmentation class imbalance terrain segmentation U-Net |
| title | FCIoU: A Targeted Approach for Improving Minority Class Detection in Semantic Segmentation Systems |
| title_full | FCIoU: A Targeted Approach for Improving Minority Class Detection in Semantic Segmentation Systems |
| title_fullStr | FCIoU: A Targeted Approach for Improving Minority Class Detection in Semantic Segmentation Systems |
| title_full_unstemmed | FCIoU: A Targeted Approach for Improving Minority Class Detection in Semantic Segmentation Systems |
| title_short | FCIoU: A Targeted Approach for Improving Minority Class Detection in Semantic Segmentation Systems |
| title_sort | fciou a targeted approach for improving minority class detection in semantic segmentation systems |
| topic | loss function off-road semantic segmentation class imbalance terrain segmentation U-Net |
| url | https://www.mdpi.com/2504-4990/5/4/85 |
| work_keys_str_mv | AT jonathanplangger fciouatargetedapproachforimprovingminorityclassdetectioninsemanticsegmentationsystems AT mohamedatia fciouatargetedapproachforimprovingminorityclassdetectioninsemanticsegmentationsystems AT hichamchaoui fciouatargetedapproachforimprovingminorityclassdetectioninsemanticsegmentationsystems |