Towards 3D Vision with Low-Cost Single-Photon Cameras

We present a method for reconstructing 3D shape of arbitrary Lambertian objects based on measurements by miniature, energy-efficient, low-cost single-photon cameras. These cameras, operating as time resolved image sensors, illuminate the scene with a very fast pulse of diffuse light and record the shape of that pulse as it returns back from the scene at a high temporal resolution. We propose to model this image formation process, account for its non-idealities, and adapt neural rendering to reconstruct 3D geometry from a set of spatially distributed sensors with known poses. We show that our approach can successfully recover complex 3D shapes from simulated data. We further demonstrate 3D object reconstruction from real-world captures, utilizing measurements from a commodity proximity sensor. Our work draws a connection between image-based modeling and active range scanning and is a step towards 3D vision with single-photon cameras.

Mitigating Fine-tuning Jailbreak Attack with Backdoor Enhanced Alignment

Despite the general capabilities of Large Language Models (LLMs) like GPT-4 and Llama-2, these models still request fine-tuning or adaptation with customized data when it comes to meeting the specific business demands and intricacies of tailored use cases. However, this process inevitably introduces new safety threats, particularly against the Fine-tuning based Jailbreak Attack (FJAttack), where incorporating just a few harmful examples into the fine-tuning dataset can significantly compromise the model safety. Though potential defenses have been proposed by incorporating safety examples into the fine-tuning dataset to reduce the safety issues, such approaches require incorporating a substantial amount of safety examples, making it inefficient. To effectively defend against the FJAttack with limited safety examples, we propose a Backdoor Enhanced Safety Alignment method inspired by an analogy with the concept of backdoor attacks. In particular, we construct prefixed safety examples by integrating a secret prompt, acting as a "backdoor trigger", that is prefixed to safety examples. Our comprehensive experiments demonstrate that through the Backdoor Enhanced Safety Alignment with adding as few as 11 prefixed safety examples, the maliciously fine-tuned LLMs will achieve similar safety performance as the original aligned models. Furthermore, we also explore the effectiveness of our method in a more practical setting where the fine-tuning data consists of both FJAttack examples and the fine-tuning task data. Our method shows great efficacy in defending against FJAttack without harming the performance of fine-tuning tasks.

A Physical-World Adversarial Attack Against 3D Face Recognition

3D face recognition systems have been widely employed in intelligent terminals, among which structured light imaging is a common method to measure the 3D shape. However, this method could be easily attacked, leading to inaccurate 3D face recognition. In this paper, we propose a novel, physically-achievable attack on the fringe structured light system, named structured light attack. The attack utilizes a projector to project optical adversarial fringes on faces to generate point clouds with well-designed noises. We firstly propose a 3D transform-invariant loss function to enhance the robustness of 3D adversarial examples in the physical-world attack. Then we reverse the 3D adversarial examples to the projector's input to place noises on phase-shift images, which models the process of structured light imaging. A real-world structured light system is constructed for the attack and several state-of-the-art 3D face recognition neural networks are tested. Experiments show that our method can attack the physical system successfully and only needs minor modifications of projected images.