DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | Kum, Dongsuk | - |
dc.contributor.advisor | 금동석 | - |
dc.contributor.author | Shin, Juyeb | - |
dc.date.accessioned | 2023-06-22T19:31:38Z | - |
dc.date.available | 2023-06-22T19:31:38Z | - |
dc.date.issued | 2023 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=1032343&flag=dissertation | en_US |
dc.identifier.uri | http://hdl.handle.net/10203/308260 | - |
dc.description | 학위논문(석사) - 한국과학기술원 : 로봇공학학제전공, 2023.2,[iv, 42 p. :] | - |
dc.description.abstract | The construction of lightweight High-definition (HD) maps containing geometric and semantic information is of foremost importance for the large-scale deployment of autonomous driving. To automatically generate such type of map from a set of images captured by a vehicle, most works formulate this mapping as a segmentation problem, which implies heavy post-processing to obtain the final vectorized representation. Alternative techniques have the ability to generate an HD map in an end-to-end manner but rely on computationally expensive auto-regressive models. To bring camera-based to an applicable level, we propose a fast end-to-end network generating a vectorized HD map via instance-level graph modeling of the map elements. Comprehensive experiments on nuScenes dataset show that our proposed network outperforms state-of-the-art methods by 13.7 mAP and achieves comparable accuracy with 5× faster inference speed. | - |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | High-definition map▼aDeep learning▼aAutonomous vehicle▼aConvolutional neural network▼aGraph neural network | - |
dc.subject | 정밀도로지도▼a심층 학습▼a자율주행 자동차▼a합성곱 신경망▼a그래프 신경망 | - |
dc.title | Instance-level graph modeling for end-to-end vectorized HD map learning | - |
dc.title.alternative | End-to-end 벡터화 정밀도로지도 학습을 위한 그래프 모델링 기법 | - |
dc.type | Thesis(Master) | - |
dc.identifier.CNRN | 325007 | - |
dc.description.department | 한국과학기술원 :로봇공학학제전공, | - |
dc.contributor.alternativeauthor | 신주엽 | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.