A Hybrid Image Segmentation Method for Accurate Measurement of Urban Environments
In the field of urban environment analysis research, image segmentation technology that groups important objects in the urban landscape image in pixel units has been the subject of increased attention. However, since a dataset consisting of a huge amount of image and label pairs is required to utili...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-04-01
|
| Series: | Electronics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2079-9292/12/8/1845 |
| _version_ | 1797605626615955456 |
|---|---|
| author | Hyungjoon Kim Jae Ho Lee Suan Lee |
| author_facet | Hyungjoon Kim Jae Ho Lee Suan Lee |
| author_sort | Hyungjoon Kim |
| collection | DOAJ |
| description | In the field of urban environment analysis research, image segmentation technology that groups important objects in the urban landscape image in pixel units has been the subject of increased attention. However, since a dataset consisting of a huge amount of image and label pairs is required to utilize this technology, in most cases, a model trained with a dataset having similar characteristics is used for analysis, and as a result, the quality of segmentation is poor. To overcome this limitation, we propose a hybrid model to leverage the strengths of each model in predicting specific classes. In particular, we first introduce a pre-processing operation to reduce the differences between the collected urban dataset and public dataset. Subsequently, we train several segmentation models with a pre-processed dataset then, based on the weight rule, the segmentation results are fused to create one segmentation map. To evaluate our proposal, we collected Google Street View images that do not have any labels and trained a model using the cityscapes dataset which contains foregrounds similar to the collected images. We quantitatively assessed its performance using the cityscapes dataset with ground truths and qualitatively evaluated the results of GSV data segmentation through user studies. Our approach outperformed existing methods and demonstrated the potential for accurate and efficient urban environment analysis using computer vision technology. |
| first_indexed | 2024-03-11T05:03:45Z |
| format | Article |
| id | doaj.art-1c88612c26a1448abbbe1ac19810b916 |
| institution | Directory Open Access Journal |
| issn | 2079-9292 |
| language | English |
| last_indexed | 2024-03-11T05:03:45Z |
| publishDate | 2023-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Electronics |
| spelling | doaj.art-1c88612c26a1448abbbe1ac19810b9162023-11-17T19:01:41ZengMDPI AGElectronics2079-92922023-04-01128184510.3390/electronics12081845A Hybrid Image Segmentation Method for Accurate Measurement of Urban EnvironmentsHyungjoon Kim0Jae Ho Lee1Suan Lee2School of Computer Science, Semyung University, Jecheon 27136, Republic of KoreaDepartment of Landscape Architecture, University of Seoul, Seoul 02504, Republic of KoreaSchool of Computer Science, Semyung University, Jecheon 27136, Republic of KoreaIn the field of urban environment analysis research, image segmentation technology that groups important objects in the urban landscape image in pixel units has been the subject of increased attention. However, since a dataset consisting of a huge amount of image and label pairs is required to utilize this technology, in most cases, a model trained with a dataset having similar characteristics is used for analysis, and as a result, the quality of segmentation is poor. To overcome this limitation, we propose a hybrid model to leverage the strengths of each model in predicting specific classes. In particular, we first introduce a pre-processing operation to reduce the differences between the collected urban dataset and public dataset. Subsequently, we train several segmentation models with a pre-processed dataset then, based on the weight rule, the segmentation results are fused to create one segmentation map. To evaluate our proposal, we collected Google Street View images that do not have any labels and trained a model using the cityscapes dataset which contains foregrounds similar to the collected images. We quantitatively assessed its performance using the cityscapes dataset with ground truths and qualitatively evaluated the results of GSV data segmentation through user studies. Our approach outperformed existing methods and demonstrated the potential for accurate and efficient urban environment analysis using computer vision technology.https://www.mdpi.com/2079-9292/12/8/1845urban environment analysisstreetscapesimage segmentationhybrid modeldeep learning |
| spellingShingle | Hyungjoon Kim Jae Ho Lee Suan Lee A Hybrid Image Segmentation Method for Accurate Measurement of Urban Environments Electronics urban environment analysis streetscapes image segmentation hybrid model deep learning |
| title | A Hybrid Image Segmentation Method for Accurate Measurement of Urban Environments |
| title_full | A Hybrid Image Segmentation Method for Accurate Measurement of Urban Environments |
| title_fullStr | A Hybrid Image Segmentation Method for Accurate Measurement of Urban Environments |
| title_full_unstemmed | A Hybrid Image Segmentation Method for Accurate Measurement of Urban Environments |
| title_short | A Hybrid Image Segmentation Method for Accurate Measurement of Urban Environments |
| title_sort | hybrid image segmentation method for accurate measurement of urban environments |
| topic | urban environment analysis streetscapes image segmentation hybrid model deep learning |
| url | https://www.mdpi.com/2079-9292/12/8/1845 |
| work_keys_str_mv | AT hyungjoonkim ahybridimagesegmentationmethodforaccuratemeasurementofurbanenvironments AT jaeholee ahybridimagesegmentationmethodforaccuratemeasurementofurbanenvironments AT suanlee ahybridimagesegmentationmethodforaccuratemeasurementofurbanenvironments AT hyungjoonkim hybridimagesegmentationmethodforaccuratemeasurementofurbanenvironments AT jaeholee hybridimagesegmentationmethodforaccuratemeasurementofurbanenvironments AT suanlee hybridimagesegmentationmethodforaccuratemeasurementofurbanenvironments |