Electrical Engineering Systems Seminar

Speaker: Osher Stamker

M.Sc. student under the supervision of Prof. Shai Avidan & Prof. Alex Liberzon

Wednesday, 13^th November 2024, at 15:30

Room 011, Kitot Building, Faculty of Engineering

HydroNeRF: NeRF for Scenes with Air and Water

Abstract

Neural Radiance Fields (NeRF) revolutionized the field of view synthesis by providing high-quality, photorealistic novel views from unstructured image collections. As a side benefit, it is possible to measure geometric properties in the scene (such as distances between 3D points) based on a depth map that is extracted from NeRF.

However, NeRF models struggle with complex light interactions, such as reflections and refractions, often resulting in inaccurate renderings. Going beyond view synthesis, accurate 3D measurements of objects submerged in liquid is also important for fluid mechanics.

In this work, we introduce HydroNeRF, an enhanced NeRF framework designed to accurately model light refraction and handle reflective scenes. Our approach leverages Snell's law to model the bending of light rays as they transition between different media (air, glass, water), and integrates a Deformation Network to render the scene.

We collected a dataset of challenging scenes that contain objects submerged in water tanks and show that HydroNeRF can be used to measure distances between 3D points more accurately than NeRF.

השתתפות בסמינר תיתן קרדיט שמיעה = עפ"י רישום שם מלא + מספר ת.ז. בדף הנוכחות שיועבר באולם במהלך הסמינר

Electrical Engineering Systems Seminar

Speaker: Dror Peri

M.Sc. student under the supervision of Prof. Avishai Wool

Wednesday, 27^th November 2024, at 15:00

Room 011, Kitot Building, Faculty of Engineering

Camera Spoofing via the in-Vehicle IP Network

Abstract

Autonomous driving and advanced driver assistance systems (ADAS) rely on cameras to control driving. In many prior approaches an attacker aiming to stop the vehicle had to send messages on the specialized and better-defended CAN bus. We suggest an easier alternative: manipulate the IP-based network communication between the camera and the ADAS logic, inject fake images of stop signs or red lights into the video stream, and let the ADAS stop the car safely. We created such an attack tool that successfully exploits the GigE Vision protocol.

Then we analyze two classes of passive anomaly detectors to identify such attacks: protocol-based detectors and video-based detectors. We implemented multiple detectors of both classes and evaluated them on data collected from our test vehicle and on data from the public BDD corpus. Our results show that such detectors are effective against naive adversaries, but sophisticated adversaries can evade detection.

Finally, we propose a novel class of active defense mechanisms that randomly adjust camera parameters during the video transmission and verify that the received images obey the requested adjustments. Within this class we focus on a specific implementation, the width-varying defense, which randomly modifies the width of every frame. Beyond its function as an anomaly detector, this defense is also a protective measure against certain attacks: by distorting injected image patches it prevents their recognition by the ADAS logic. We demonstrate the effectiveness of the width-varying defense through theoretical analysis and by an extensive evaluation of several types of attack in a wide range of realistic road driving conditions. The best the attack was able to achieve against this defense was injecting a stop sign for a duration of 0.2 seconds, with a success probability of 0.2%, whereas stopping a vehicle requires about 2.5 seconds.

דןעהשתתפות בסמינר תיתן קרדיט שמיעה = עפ"י רישום שם מלא + מספר ת.ז. בדף הנוכחות שיועבר באולם במהלך הסמינר