Deepfake detectors have demonstrated vulnerability when faced with adversarial attacks. In real-world scenarios, attackers can generate adversarial examples that mislead detectors through query-based black-box attack methods. Recently, several defense methods against query-based adversarial attacks have achieved promising performance. However, their effectiveness significantly declines when applied to deepfake detection tasks. In this work, we propose a novel defense method specifically designed to counter adversarial attacks in deepfake detection scenarios. Unlike existing approaches, our method focuses on the high-frequency components of images to uncover subtle traces left by adversarial perturbations. Specifically, we analyze the high-frequency residual between similar images within the queries to detect adversarial examples. We evaluate our method against three advanced black-box attack strategies. Extensive experiments demonstrate that our approach achieves highly effective attack detection while significantly reducing FPR.

The rapid advancement of generative models has led to the widespread emergence of highly realistic synthetic images, making the detection of AI-generated content increasingly critical. In particular, diffusion models have recently achieved unprecedented levels of visual fidelity, further raising concerns. While most existing approaches rely on supervised learning, zero-shot detection methods have attracted growing interest due to their ability to bypass data collection and maintenance. Nevertheless, the performance of current zero-shot methods remains limited. In this paper, we introduce a novel zero-shot AI-generated image detection method. Unlike previous works that primarily focus on identifying artifacts in the final generated images, our work explores features within the image generation process that can be leveraged for detection. Specifically, we simulate the image sampling process via diffusion-based inversion and observe that the denoising outputs of generated images converge to the target image more rapidly than those of real images. Inspired by this observation, we compute the similarity between the original image and the outputs along the denoising trajectory, which is then used as an indicator of image authenticity.Since our method requires no training on any generated images, it avoids overfitting to specific generative models or dataset biases. Experiments across a wide range of generators demonstrate that our method achieves significant improvements over state-of-the-art supervised and zero-shot counterparts.

Detection of face forgery videos remains a formidable challenge in the field of digital forensics, especially the generalization to unseen datasets and common perturbations. In this paper, we tackle this issue by leveraging the synergy between audio and visual speech elements, embarking on a novel approach through audio-visual speech representation learning. Our work is motivated by the finding that audio signals, enriched with speech content, can provide precise information effectively reflecting facial movements. To this end, we first learn precise audio-visual speech representations on real videos via a self-supervised masked prediction task, which encodes both local and global semantic information simultaneously. Then, the derived model is directly transferred to the forgery detection task. Extensive experiments demonstrate that our method outperforms the state-of-the-art methods in terms of cross-dataset generalization and robustness, without the participation of any fake video in model training.

The interface between the Radio Access Network and Core Network is kept for Radio Resource Control (RRC) Inactive relative to RRC Idle, and the initial access should be conducted for inactive user equipments (UEs) to reduce the delay of link connection when inactive UEs switch to active. However, due to less interaction information between inactive UEs and transmission and reception points (TRPs), it is necessary to conduct initial access by beam sweeping rather than refined beam alignment. In this paper, an effective uplink multi-beam initial access scheme for inactive UEs in mmWave networks is proposed, where uplink beam sweeping and fast beam recovery is considered. Due to the application of digital beamforming technology and considering that TRP scanning beam number is more than that of UE scanning beams, the delay of uplink based beam sweeping is significantly reduced in the proposed initial access scheme, compared with the traditional downlink based beam sweeping. Furthermore, when the serving beam pair is blocked or not available, fast beam recovery would be conducted by using backup beam pair, where the cases that the backup beam pair and serving beam pair belonging to the same TRP or different TRPs are considered. Simulation results present the effect of density of TRPs, beam sweeping cycle and UE speed on outage probability and average rate after inactive UEs switching to active, and verify initial access performance gain of the proposed initial access scheme.

Ultra-Dense Network (UDN) is a clear trend for future 5G. However, using exciting small cell selection (SCS) schemes, which are totally based on downlink measurement, dense deployment of small cells would bring about frequent measurement, intolerable ping-pong (PP) rate and handover failure (HOF) rate. In this paper, an uplink measurement based SCS scheme is proposed for a user-centric networks based on the analysis of mobility management challenges of UDN, and the handover performance of the proposed SCS scheme is evaluated. In the proposed SCS scheme, the small cells measure uplink reference signals of users and then user-centric new radio (NR) cells would be established, which reduces the measurement process that the user feedbacks measurement reports to small cells that have transmitted downlink reference signals to the user. The relationship between key handover parameters and mobility performance in user-centric network is illustrated to look for the optimal configuration of Time-to-Trigger (TTT) and A3 Offset and achieve low PP rate and HOF rate. Besides, the impact of NR cell size on handover performance is evaluated to provide basis for the forming/reforming of user-centric NR cell. Simulation results show that compared with traditional downlink measurement based SCS scheme, the proposed SCS scheme can reduce the PP rate and HOF rate by up to 21% and 56%, respectively.