Non-Anchor-Based Vehicle Detection for Traffic Surveillance Using Bounding Ellipses
Cameras for traffic surveillance are usually pole-mounted and produce images that reflect a birds-eye view. Vehicles in such images, in general, assume an ellipse form. A bounding box for the vehicles usually includes a large empty space when the vehicle orientation is not parallel to the edges of t...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2021-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/9526639/ |
| _version_ | 1818912292636459008 |
|---|---|
| author | Byeonghyeop Yu Johyun Shin Gyeongjun Kim Seungbin Roh Keemin Sohn |
| author_facet | Byeonghyeop Yu Johyun Shin Gyeongjun Kim Seungbin Roh Keemin Sohn |
| author_sort | Byeonghyeop Yu |
| collection | DOAJ |
| description | Cameras for traffic surveillance are usually pole-mounted and produce images that reflect a birds-eye view. Vehicles in such images, in general, assume an ellipse form. A bounding box for the vehicles usually includes a large empty space when the vehicle orientation is not parallel to the edges of the box. To circumvent this problem, the present study applied bounding ellipses to a non-anchor-based, single-shot detection model (CenterNet). Since this model does not depend on anchor boxes, non-max suppression (NMS) that requires computing the intersection over union (IOU) between predicted bounding boxes is unnecessary for inference. The SpotNet that extends the CenterNet model by adding a segmentation head was also tested with bounding ellipses. Two other anchor-based, single-shot detection models (YOLO4 and SSD) were chosen as references for comparison. The model performance was compared based on a local dataset that was doubly annotated with bounding boxes and ellipses. As a result, the performance of the two models with bounding ellipses exceeded that of the reference models with bounding boxes. When the backbone of the ellipse models was pretrained on an open dataset (UA-DETRAC), the performance was further enhanced. Several data augmentation schemes also improved the performance of the proposed models. As a result, the best mAP score of a CenterNet exceeds 0.95 when augmenting heatmaps with bounding ellipses. |
| first_indexed | 2024-12-19T23:12:17Z |
| format | Article |
| id | doaj.art-fbbae66c1f99475a953e85bbcf7c79b5 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-12-19T23:12:17Z |
| publishDate | 2021-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-fbbae66c1f99475a953e85bbcf7c79b52022-12-21T20:02:11ZengIEEEIEEE Access2169-35362021-01-01912306112307410.1109/ACCESS.2021.31092589526639Non-Anchor-Based Vehicle Detection for Traffic Surveillance Using Bounding EllipsesByeonghyeop Yu0Johyun Shin1https://orcid.org/0000-0001-5057-3446Gyeongjun Kim2Seungbin Roh3https://orcid.org/0000-0001-7705-0282Keemin Sohn4https://orcid.org/0000-0002-7270-7094Department of Urban Engineering, Chung-Ang University, Seoul, South KoreaDepartment of Urban Engineering, Chung-Ang University, Seoul, South KoreaDepartment of Urban Engineering, Chung-Ang University, Seoul, South KoreaDepartment of Urban Engineering, Chung-Ang University, Seoul, South KoreaDepartment of Urban Engineering, Department of Smart Cities, Chung-Ang University, Seoul, South KoreaCameras for traffic surveillance are usually pole-mounted and produce images that reflect a birds-eye view. Vehicles in such images, in general, assume an ellipse form. A bounding box for the vehicles usually includes a large empty space when the vehicle orientation is not parallel to the edges of the box. To circumvent this problem, the present study applied bounding ellipses to a non-anchor-based, single-shot detection model (CenterNet). Since this model does not depend on anchor boxes, non-max suppression (NMS) that requires computing the intersection over union (IOU) between predicted bounding boxes is unnecessary for inference. The SpotNet that extends the CenterNet model by adding a segmentation head was also tested with bounding ellipses. Two other anchor-based, single-shot detection models (YOLO4 and SSD) were chosen as references for comparison. The model performance was compared based on a local dataset that was doubly annotated with bounding boxes and ellipses. As a result, the performance of the two models with bounding ellipses exceeded that of the reference models with bounding boxes. When the backbone of the ellipse models was pretrained on an open dataset (UA-DETRAC), the performance was further enhanced. Several data augmentation schemes also improved the performance of the proposed models. As a result, the best mAP score of a CenterNet exceeds 0.95 when augmenting heatmaps with bounding ellipses.https://ieeexplore.ieee.org/document/9526639/Bounding ellipsedeep-learningtraffic surveillanceobjects as pointsvehicle detection |
| spellingShingle | Byeonghyeop Yu Johyun Shin Gyeongjun Kim Seungbin Roh Keemin Sohn Non-Anchor-Based Vehicle Detection for Traffic Surveillance Using Bounding Ellipses IEEE Access Bounding ellipse deep-learning traffic surveillance objects as points vehicle detection |
| title | Non-Anchor-Based Vehicle Detection for Traffic Surveillance Using Bounding Ellipses |
| title_full | Non-Anchor-Based Vehicle Detection for Traffic Surveillance Using Bounding Ellipses |
| title_fullStr | Non-Anchor-Based Vehicle Detection for Traffic Surveillance Using Bounding Ellipses |
| title_full_unstemmed | Non-Anchor-Based Vehicle Detection for Traffic Surveillance Using Bounding Ellipses |
| title_short | Non-Anchor-Based Vehicle Detection for Traffic Surveillance Using Bounding Ellipses |
| title_sort | non anchor based vehicle detection for traffic surveillance using bounding ellipses |
| topic | Bounding ellipse deep-learning traffic surveillance objects as points vehicle detection |
| url | https://ieeexplore.ieee.org/document/9526639/ |
| work_keys_str_mv | AT byeonghyeopyu nonanchorbasedvehicledetectionfortrafficsurveillanceusingboundingellipses AT johyunshin nonanchorbasedvehicledetectionfortrafficsurveillanceusingboundingellipses AT gyeongjunkim nonanchorbasedvehicledetectionfortrafficsurveillanceusingboundingellipses AT seungbinroh nonanchorbasedvehicledetectionfortrafficsurveillanceusingboundingellipses AT keeminsohn nonanchorbasedvehicledetectionfortrafficsurveillanceusingboundingellipses |