• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI

    Davide Scaramuzza Releases Hybrid Event Camera Simulator

    HESIM is presented as the first hybrid event camera simulator developed with AlpsenTek.

    KD
    DS
    2 Sources, 30d ago, first seen 30d ago

    TLDR

    Professor Davide Scaramuzza announced the release of HESIM from his Robotics and Perception Group at the University of Zurich. The project is described as the first Hybrid Event Camera Simulator and stems from direct collaboration with AlpsenTek. Scaramuzza shared the news in a public post that was later retweeted by academic Kosta Derpanis. He referenced the simulator during a recent keynote. The release targets researchers working on event cameras and related perception systems.

    Combined views

    3.9K

    2 Sources, first seen 30d ago

    Combined views

    3.9K

    2 Sources, first seen 30d ago

    66 likes
    66 likes
    3 comments
    38 saves
    18 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    3 comments
    38 saves
    18 reposts

    Sentiment

    Positive——Negative

    Summary

    Not enough discussion yet.

    No sentiment analysis available yet.

    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    2 Sources

    @davsca1We are excited to release #HESIM, the first Hybrid Event Camera Simulator, done in collaboration with AlpsenTek, to be presented at #ECCV2026 on Thu, Sep 10, from 7:30 AM – 9:30 AM in ExHall #271. PDF: https://arxiv.org/pdf/2511.18037 Code: https://github.com/yunfanLu/HESIM As mentioned in my last keynote (min 15:28, https://www.youtube.com/watch?v=A-m_1UQ-sbI&t=928s), in my group, we believe the future of vision and robotics will feature the combination of event cameras and standard cameras, as the two sensing modalities are highly complementary: event cameras provide extremely high temporal resolution but relatively low spatial resolution, while standard cameras provide high spatial resolution and image quality but relatively low temporal resolution. So far, combining the two modalities has required compromises: placing two sensors side by side introduces spatial disparity; beam splitters avoid disparity but divide the incoming light and add optical complexity; and DAVIS cameras integrate events and frames in the same pixel array, but with an architecture primarily optimized for event sensing, resulting in compromises in conventional RGB image quality and resolution. A new generation of hybrid event cameras is now emerging (see SONY, Omnivision, AlpsenTek). Rather than treating frames as an auxiliary modality, these sensors are designed to provide high-quality conventional RGB imaging together with high-temporal-resolution event information in the same sensor. AlpsenTek's Hybrid Vision Sensor takes an interesting approach: standard RGB pixels and event-sensing pixels are integrated into the same pixel array using a Quad-Bayer arrangement. This provides spatially aligned RGB and event information from a single compact sensor, avoiding the calibration, disparity, and optical losses of traditional two-camera solutions. However, a new sensor also creates a new challenge: how do we develop and train algorithms for it before large datasets and hardware are widely available? This is why we developed HESIM, a simulator for hybrid event cameras. HESIM enables researchers to generate realistic, spatially aligned RGB and event data and develop algorithms exploiting both modalities. We hope that releasing HESIM will help the community explore this emerging sensing paradigm and accelerate research on hybrid event-based vision. Curious to work on event cameras? Apply for a PhD or Postdoc position and join us! https://rpg.ifi.uzh.ch/positions.html Come talk to us at #ECCV2026 on Thu, Sep 10, from 7:30 AM – 9:30 AM in ExHall #271. Kudos to Yunfan LU, Nico Messikommer, Nikola Zubić! Reference: Yunfan Lu, Nico Messikommer, Xiaogang Xu, Liming Chen, Yuhan Chen, Nikola Zubić, Davide Scaramuzza, Hui Xiong Hybrid Event Frame Sensors: Modeling, Calibration, and Simulation European Conference on Computer Vision, 2026
    @CSProfKGDRT @davsca1: We are excited to release #HESIM, the first Hybrid Event Camera Simulator, done in collaboration with AlpsenTek, to be present…

    2 Sources

    @davsca1We are excited to release #HESIM, the first Hybrid Event Camera Simulator, done in collaboration with AlpsenTek, to be presented at #ECCV2026 on Thu, Sep 10, from 7:30 AM – 9:30 AM in ExHall #271. PDF: https://arxiv.org/pdf/2511.18037 Code: https://github.com/yunfanLu/HESIM As mentioned in my last keynote (min 15:28, https://www.youtube.com/watch?v=A-m_1UQ-sbI&t=928s), in my group, we believe the future of vision and robotics will feature the combination of event cameras and standard cameras, as the two sensing modalities are highly complementary: event cameras provide extremely high temporal resolution but relatively low spatial resolution, while standard cameras provide high spatial resolution and image quality but relatively low temporal resolution. So far, combining the two modalities has required compromises: placing two sensors side by side introduces spatial disparity; beam splitters avoid disparity but divide the incoming light and add optical complexity; and DAVIS cameras integrate events and frames in the same pixel array, but with an architecture primarily optimized for event sensing, resulting in compromises in conventional RGB image quality and resolution. A new generation of hybrid event cameras is now emerging (see SONY, Omnivision, AlpsenTek). Rather than treating frames as an auxiliary modality, these sensors are designed to provide high-quality conventional RGB imaging together with high-temporal-resolution event information in the same sensor. AlpsenTek's Hybrid Vision Sensor takes an interesting approach: standard RGB pixels and event-sensing pixels are integrated into the same pixel array using a Quad-Bayer arrangement. This provides spatially aligned RGB and event information from a single compact sensor, avoiding the calibration, disparity, and optical losses of traditional two-camera solutions. However, a new sensor also creates a new challenge: how do we develop and train algorithms for it before large datasets and hardware are widely available? This is why we developed HESIM, a simulator for hybrid event cameras. HESIM enables researchers to generate realistic, spatially aligned RGB and event data and develop algorithms exploiting both modalities. We hope that releasing HESIM will help the community explore this emerging sensing paradigm and accelerate research on hybrid event-based vision. Curious to work on event cameras? Apply for a PhD or Postdoc position and join us! https://rpg.ifi.uzh.ch/positions.html Come talk to us at #ECCV2026 on Thu, Sep 10, from 7:30 AM – 9:30 AM in ExHall #271. Kudos to Yunfan LU, Nico Messikommer, Nikola Zubić! Reference: Yunfan Lu, Nico Messikommer, Xiaogang Xu, Liming Chen, Yuhan Chen, Nikola Zubić, Davide Scaramuzza, Hui Xiong Hybrid Event Frame Sensors: Modeling, Calibration, and Simulation European Conference on Computer Vision, 2026
    @CSProfKGDRT @davsca1: We are excited to release #HESIM, the first Hybrid Event Camera Simulator, done in collaboration with AlpsenTek, to be present…