3D Vehicle Detection and Segmentation Based on EfficientNetB3 and CenterNet Residual Blocks
In this paper, we present a two stages solution to 3D vehicle detection and segmentation. The first stage depends on the combination of EfficientNetB3 architecture with multiparallel residual blocks (inspired by CenterNet architecture) for 3D localization and poses estimation for vehicles on the sce...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-10-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/22/20/7990 |
| _version_ | 1797469923906158592 |
|---|---|
| author | Alexey Kashevnik Ammar Ali |
| author_facet | Alexey Kashevnik Ammar Ali |
| author_sort | Alexey Kashevnik |
| collection | DOAJ |
| description | In this paper, we present a two stages solution to 3D vehicle detection and segmentation. The first stage depends on the combination of EfficientNetB3 architecture with multiparallel residual blocks (inspired by CenterNet architecture) for 3D localization and poses estimation for vehicles on the scene. The second stage takes the output of the first stage as input (cropped car images) to train EfficientNet B3 for the image recognition task. Using predefined 3D Models, we substitute each vehicle on the scene with its match using the rotation matrix and translation vector from the first stage to get the 3D detection bounding boxes and segmentation masks. We trained our models on an open-source dataset (ApolloCar3D). Our method outperforms all published solutions in terms of 6 degrees of freedom error (6 DoF err). |
| first_indexed | 2024-03-09T19:29:50Z |
| format | Article |
| id | doaj.art-8a994f0ac90f4530bbdd1adf56954200 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-09T19:29:50Z |
| publishDate | 2022-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-8a994f0ac90f4530bbdd1adf569542002023-11-24T02:29:55ZengMDPI AGSensors1424-82202022-10-012220799010.3390/s222079903D Vehicle Detection and Segmentation Based on EfficientNetB3 and CenterNet Residual BlocksAlexey Kashevnik0Ammar Ali1St. Petersburg Federal Research Center of the Russian Academy of Sciences, SPC RAS, 199178 St. Petersburg, RussiaInformation Technology and Programming Faculty, ITMO University, 197101 St. Petersburg, RussiaIn this paper, we present a two stages solution to 3D vehicle detection and segmentation. The first stage depends on the combination of EfficientNetB3 architecture with multiparallel residual blocks (inspired by CenterNet architecture) for 3D localization and poses estimation for vehicles on the scene. The second stage takes the output of the first stage as input (cropped car images) to train EfficientNet B3 for the image recognition task. Using predefined 3D Models, we substitute each vehicle on the scene with its match using the rotation matrix and translation vector from the first stage to get the 3D detection bounding boxes and segmentation masks. We trained our models on an open-source dataset (ApolloCar3D). Our method outperforms all published solutions in terms of 6 degrees of freedom error (6 DoF err).https://www.mdpi.com/1424-8220/22/20/7990autonomous driving3D object detectionlocalizationimage processingmachine learningvehicle classification |
| spellingShingle | Alexey Kashevnik Ammar Ali 3D Vehicle Detection and Segmentation Based on EfficientNetB3 and CenterNet Residual Blocks Sensors autonomous driving 3D object detection localization image processing machine learning vehicle classification |
| title | 3D Vehicle Detection and Segmentation Based on EfficientNetB3 and CenterNet Residual Blocks |
| title_full | 3D Vehicle Detection and Segmentation Based on EfficientNetB3 and CenterNet Residual Blocks |
| title_fullStr | 3D Vehicle Detection and Segmentation Based on EfficientNetB3 and CenterNet Residual Blocks |
| title_full_unstemmed | 3D Vehicle Detection and Segmentation Based on EfficientNetB3 and CenterNet Residual Blocks |
| title_short | 3D Vehicle Detection and Segmentation Based on EfficientNetB3 and CenterNet Residual Blocks |
| title_sort | 3d vehicle detection and segmentation based on efficientnetb3 and centernet residual blocks |
| topic | autonomous driving 3D object detection localization image processing machine learning vehicle classification |
| url | https://www.mdpi.com/1424-8220/22/20/7990 |
| work_keys_str_mv | AT alexeykashevnik 3dvehicledetectionandsegmentationbasedonefficientnetb3andcenternetresidualblocks AT ammarali 3dvehicledetectionandsegmentationbasedonefficientnetb3andcenternetresidualblocks |