DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | Je, Minkyu | - |
dc.contributor.advisor | 제민규 | - |
dc.contributor.author | Fatima, Mishal | - |
dc.date.accessioned | 2021-05-13T19:39:42Z | - |
dc.date.available | 2021-05-13T19:39:42Z | - |
dc.date.issued | 2020 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=925240&flag=dissertation | en_US |
dc.identifier.uri | http://hdl.handle.net/10203/285076 | - |
dc.description | 학위논문(석사) - 한국과학기술원 : 전기및전자공학부, 2020.8,[iv, 33 p. :] | - |
dc.description.abstract | Anticipation of accidents ahead of time in autonomous and non-autonomous vehicles aids in accident avoidance. In order to recognize abnormal events such as traffic accidents in a video sequence, it is important that the network takes into account interactions of objects in a given frame. We propose a novel Feature Aggregation (FA) block that has two main functions. First, it computes appearance relation between different objects in a single frame and gives rise to attention weights. Afterwards, each object’s features are refined by computing a weighted sum of the features of all objects in a frame whereby the weights are obtained from the first step. We use FA block along with Long Short Term Memory (LSTM) network to anticipate accidents in the video sequences. We report mean Average Precision (mAP) and Average Time-to-Accident (ATTA) on Street Accident (SA) dataset. Our proposed method achieves the highest score for risk anticipation by predicting accidents 0.32 sec and 0.75 sec earlier compared to the best results with Adaptive Loss and dynamic parameter prediction based methods respectively. | - |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | Inter-object interaction▼aappearance comparison▼aactivity anticipation▼afeature aggregation▼aLSTM | - |
dc.subject | 개체 간 상호 작용 | - |
dc.title | Global feature aggregation for accident anticipation | - |
dc.title.alternative | 사고 예측을 위한 전역적 특징의 집계 처리 기법 | - |
dc.type | Thesis(Master) | - |
dc.identifier.CNRN | 325007 | - |
dc.description.department | 한국과학기술원 :전기및전자공학부, | - |
dc.contributor.alternativeauthor | 파티마미쌀 | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.