ISSAFE: Improving Semantic Segmentation in Accidents by Fusing Event-based Data

被引:22
|
作者
Zhang, Jiaming [1 ]
Yang, Kailun [1 ]
Stiefelhagen, Rainer [1 ]
机构
[1] Karlsruhe Inst Technol, Inst Anthropomat & Robot, Karlsruhe, Germany
来源
2021 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS (IROS) | 2021年
关键词
D O I
10.1109/IROS51168.2021.9636109
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Ensuring the safety of all traffic participants is a prerequisite for bringing intelligent vehicles closer to practical applications. The assistance system should not only achieve high accuracy under normal conditions, but obtain robust perception against extreme situations. However, traffic accidents that involve object collisions, deformations, overturns, etc., yet unseen in most training sets, will largely harm the performance of existing semantic segmentation models. To tackle this issue, we present a rarely addressed task regarding semantic segmentation in accidental scenarios, along with an accident dataset DADA-seg. It contains 313 various accident sequences with 40 frames each, of which the time windows are located before and during a traffic accident. Every 11th frame is manually annotated for benchmarking the segmentation performance. Furthermore, we propose a novel event-based multi-modal segmentation architecture ISSAFE. Our experiments indicate that event-based data can provide complementary information to stabilize semantic segmentation under adverse conditions by preserving fine-grain motion of fast-moving foreground (crash objects) in accidents. Our approach achieves +8.2% mIoU performance gain on the proposed evaluation set, exceeding more than 10 state-of-the-art segmentation methods. The proposed ISSAFE architecture is demonstrated to be consistently effective for models learned on multiple source databases including Cityscapes, KITTI-360, BDD and ApolloScape.
引用
收藏
页码:1132 / 1139
页数:8
相关论文
共 50 条
  • [21] Event-based Sport Video Segmentation using Multimodal Analysis
    Hahm, Gyeong-June
    Cho, Keeseong
    2016 INTERNATIONAL CONFERENCE ON INFORMATION AND COMMUNICATION TECHNOLOGY CONVERGENCE (ICTC 2016): TOWARDS SMARTER HYPER-CONNECTED WORLD, 2016, : 1119 - 1121
  • [22] Event-based compression and mining of data streams
    Cuzzocrea, Alfredo
    Chakravarthy, Sharma
    KNOWLEDGE-BASED INTELLIGENT INFORMATION AND ENGINEERING SYSTEMS, PT 2, PROCEEDINGS, 2008, 5178 : 670 - +
  • [23] Event-based processing of neutron scattering data
    Peterson, Peter F.
    Campbell, Stuart I.
    Reuter, Michael A.
    Taylor, Russell J.
    Zikovsky, Janik
    NUCLEAR INSTRUMENTS & METHODS IN PHYSICS RESEARCH SECTION A-ACCELERATORS SPECTROMETERS DETECTORS AND ASSOCIATED EQUIPMENT, 2015, 803 : 24 - 28
  • [24] Advances in an Event-Based Spatiotemporal Data Modeling
    Zhu, Xinming
    Liu, Haiyan
    Xu, Qing
    Liu, Jun'nan
    Lihua, Xiaoyang
    SCIENTIFIC PROGRAMMING, 2021, 2021
  • [25] Event-Based Color Segmentation With a High Dynamic Range Sensor
    Marcireau, Alexandre
    Ieng, Sio-Hoi
    Simon-Chane, Camille
    Benosman, Ryad B.
    FRONTIERS IN NEUROSCIENCE, 2018, 12
  • [26] EventASEG: An Event-Based Asynchronous Segmentation of Road With Likelihood Attention
    Annamalai, Lakshmi
    Thakur, Chetan Singh
    IEEE ROBOTICS AND AUTOMATION LETTERS, 2024, 9 (08): : 6951 - 6958
  • [27] Event-based Organization Model for Sensing Data
    Sun, Yunchuan
    Zhang, Junsheng
    Bie, Rongfang
    Yan, Hongli
    Xia, Ye
    Zhou, Zhangbing
    2014 IEEE COMPUTING, COMMUNICATIONS AND IT APPLICATIONS CONFERENCE (COMCOMAP), 2014, : 239 - 242
  • [28] Adversarial Attack for Asynchronous Event-Based Data
    Lee, Wooju
    Myung, Hyun
    THIRTY-SIXTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE / THIRTY-FOURTH CONFERENCE ON INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE / THE TWELVETH SYMPOSIUM ON EDUCATIONAL ADVANCES IN ARTIFICIAL INTELLIGENCE, 2022, : 1237 - 1244
  • [29] EventDrop: Data Augmentation for Event-based Learning
    Gu, Fuqiang
    Sng, Weicong
    Hu, Xuke
    Yu, Fangwen
    PROCEEDINGS OF THE THIRTIETH INTERNATIONAL JOINT CONFERENCE ON ARTIFICIAL INTELLIGENCE, IJCAI 2021, 2021, : 700 - 707
  • [30] ProgressiveMotionSeg: Mutually Reinforced Framework for Event-Based Motion Segmentation
    Chen, Jinze
    Wang, Yang
    Cao, Yang
    Wu, Feng
    Zha, Zheng-Jun
    THIRTY-SIXTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE / THIRTY-FOURTH CONFERENCE ON INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE / THE TWELVETH SYMPOSIUM ON EDUCATIONAL ADVANCES IN ARTIFICIAL INTELLIGENCE, 2022, : 303 - 311