Image retouching aims to enhance visual quality while aligning with users' personalized aesthetic preferences. To address the challenge of balancing controllability and subjectivity, we propose a unified diffusion-based image retouching framework called PerTouch. Our method supports semantic-level image retouching while maintaining global aesthetics. Using parameter maps containing attribute values in specific semantic regions as input, PerTouch constructs an explicit parameter-to-image mapping for fine-grained image retouching. To improve semantic boundary perception, we introduce semantic replacement and parameter perturbation mechanisms during training. To connect natural language instructions with visual control, we develop a VLM-driven agent to handle both strong and weak user instructions. Equipped with mechanisms of feedback-driven rethinking and scene-aware memory, PerTouch better aligns with user intent and captures long-term preferences. Extensive experiments demonstrate each component’s effectiveness and the superior performance of PerTouch in personalized image retouching.

Blind face restoration is a highly ill-posed problem due to the lack of necessary context. Although existing methods produce high-quality outputs, they often fail to faithfully preserve the individual's identity. In this paper, we propose a personalized face restoration method, FaceMe, based on a diffusion model. Given a single or a few reference images, we use an identity encoder to extract identity-related features, which serve as prompts to guide the diffusion model in restoring high-quality and identity-consistent facial images. By simply combining identity-related features, we effectively minimize the impact of identity-irrelevant features during training and support any number of reference image inputs during inference. Additionally, thanks to the robustness of the identity encoder, synthesized images can be used as reference images during training, and identity changing during inference does not require fine-tuning the model. We also propose a pipeline for constructing a reference image training pool that simulates the poses and expressions that may appear in real-world scenarios. Experimental results demonstrate that our FaceMe can restore high-quality facial images while maintaining identity consistency, achieving excellent performance and robustness.

We propose a novel Iterative Predictor-Critic Code Decoding framework for real-world image dehazing, abbreviated as IPC-Dehaze, which leverages the high-quality codebook prior encapsulated in a pre-trained VQGAN. Apart from previous codebook-based methods that rely on oneshot decoding, our method utilizes high-quality codes obtained in the previous iteration to guide the prediction of the Code-Predictor in the subsequent iteration, improving code prediction accuracy and ensuring stable dehazing performance. Our idea stems from the observations that 1) the degradation of hazy images varies with haze density and scene depth, and 2) clear regions play crucial cues in restoring dense haze regions. However, it is nontrivial to progressively refine the obtained codes in subsequent iterations, owing to the difficulty in determining which codes should be retained or replaced at each iteration. Another key insight of our study is to propose CodeCritic to capture interrelations among codes. The CodeCritic is used to evaluate code correlations and then resample a set of codes with the highest mask scores, i.e., a higher score indicates that the code is more likely to be rejected, which helps retain more accurate codes and predict difficult ones. Extensive experiments demonstrate the superiority of our method over state-of-the-art methods in real-world dehazing. Our project page can be found at https://github.com/Jiayi-Fu/IPC-Dehaze.

The Vector Quantized Variational AutoEncoder (VQ-VAE) has shown great potential in image generation, especially the methods with hierarchical features. However, the lack of decoupling of structural information between hierarchical features leads to semantic inconsistencies and redundant structural features, resulting in incompatible outputs. In this study, we propose the Adaptively Hierarchical Quantization Variational AutoEncoder (AHQ-VAE) to generate high-fidelity images with a unified structure. To ensure the semantic consistency of continuous space, we employ the Spatially Consistent Semantic Embedding (SCSE) module to align the hierarchical features, while decoupling global structural information and local details. To ensure the consistency of discrete space, we introduce the Adaptive Bottom Quantizer (ABQ) to generate the quantized bottom codes consistent with quantized top codes, so that the local details can adapt to the global semantics. Extensive experiments demonstrate our approach can generate high-quality images with a unified structure.

Within the domain of blind face restoration (BFR), approaches lacking facial priors frequently result in excessively smoothed visual outputs. Exiting BFR methods predominantly utilize generative facial priors to achieve realistic and authentic details. However, these methods, primarily designed for images, encounter challenges in maintaining temporal consistency when applied to face video restoration. To tackle this issue, we introduce StableBFVR, an innovative Blind Face Video Restoration method based on Stable Diffusion that incorporates temporal information into the generative prior. This is achieved through the introduction of temporal layers in the diffusion process. These temporal layers consider both long-term and short-term information aggregation. Moreover, to improve generalizability, BFR methods employ complex, large-scale degradation during training, but it often sacrifices accuracy. Addressing this, StableBFVR features a novel mixed-degradation-aware prompt module, capable of encoding specific degradation information to dynamically steer the restoration process. Comprehensive experiments demonstrate that our proposed StableBFVR outperforms state-of-the-art methods.

Image restoration has experienced significant advancements due to the development of deep learning. Nevertheless, it encounters challenges related to ill-posed problems, resulting in deviations between single model predictions and ground-truths. Ensemble learning, as a powerful machine learning technique, aims to address these deviations by combining the predictions of multiple base models. Most existing works adopt ensemble learning during the design of restoration models, while only limited research focuses on the inference-stage ensemble of pre-trained restoration models. Regression-based methods fail to enable efficient inference, leading researchers in academia and industry to prefer averaging as their choice for post-training ensemble. To address this, we reformulate the ensemble problem of image restoration into Gaussian mixture models (GMMs) and employ an expectation maximization (EM)-based algorithm to estimate ensemble weights for aggregating prediction candidates. We estimate the range-wise ensemble weights on a reference set and store them in a lookup table (LUT) for efficient ensemble inference on the test set. Our algorithm is model-agnostic and training-free, allowing seamless integration and enhancement of various pre-trained image restoration models. It consistently outperforms regression-based methods and averaging ensemble approaches on 14 benchmarks across 3 image restoration tasks, including super-resolution, deblurring and deraining. The codes and all estimated weights have been released in Github.

Abstract Indoor GPS location estimation encounters accuracy challenges from intricate building structures and diverse signal interferences. Trilateration methods utilising APs are typically employed to estimate indoor locations. Nevertheless, estimation errors from multipath effects and high power consumption of sensors employed in location estimation curtail battery life. To address this issue, research into location estimation methods utilising machine learning has been conducted. However, challenges involving the selection of the optimal access point locations and obtaining dense RSSI data have been noted. In this article presents a solution based on sparse radio maps for decreasing the expenses of collecting RSSI data while simultaneously enhancing indoor location accuracy through the integration of image data. The proposed approach integrates matrix‐based RSSI indoor positioning (M‐RIP) for initial location estimation and feature‐based image indoor positioning (F‐IIP) for position determination via image feature matching. Furthermore, extended area‐based post‐processing (EA‐PP) is employed to augment M‐RIP's precision and minimize image matching computation in F‐IIP, improving overall performance. This article utilizes actual building data to validate the precision of the position estimation and efficiency of computation reduction using the proposed method.

Abstract As more vehicles are being connected to the Internet and equipped with autonomous driving features, more robust safety and security measures are required for connected and autonomous vehicles (CAVs). Therefore, threat analysis and risk assessment are essential to prepare against cybersecurity risks for CAVs. Although prior studies have measured the possibility of attack and damage from attack as risk assessment indices, they have not analyzed the expanding attack surface or risk assessment indices that rely upon real-time resilience. This study proposes the PIER method to evaluate the cybersecurity risks of CAVs. We implemented cyber resilience for CAVs by presenting new criteria, such as exposure and recovery, in addition to probability and impact, as indices for the threat analysis and risk assessment of vehicles. To verify its effectiveness, the PIER method was evaluated with respect to software update over-the-air and collision avoidance features. Furthermore, we found that implementing security requirements that mitigate serious risks successfully diminishes the risk indices. Using the risk assessment matrix, the PIER method can shorten the risk determination time through high-risk coverage and a simple process.

In wireless ad hoc networks, it is difficult to share the information on neighbor devices in a distributed manner. Therefore, efficient neighbor discovery algorithms should be devised for self-organization in wireless ad hoc networks. In this paper, we propose a probabilistic neighbor discovery (PND) algorithm, which aims at reducing the neighbor discovery time by adjusting the transmission probability of advertisement messages through the muiltiplicative-increase/multiplicative-decrease (MIMD) policy. To further improve PND, we consider the collision detection (CD) capability in which a device can distinguish between successful reception and collision of advertisement messages. Simulation results show that the transmission probabilities of PND and PND with CD converge on the optimal value quickly although the number of devices is unknown. As a result, PND and PND with CD can reduce the neighbor discovery time by 15.6% and 57.0%, respectively, compared with the ALOHA-like neighbor discovery algorithm.

Quality of Experience (QoE) is the key to success for multimedia applications and perceptual video quality is one of the important component of QoE. A recent video encoding scheme called Scalable Video Coding (SVC) provides the flexibility and the capability to adapt the video quality to varying network conditions and heterogeneous users. In this paper, we focus on SVC multicast over IEEE 802.11 networks. Traditionally, multicast uses the lowest modulation resulting in a video with only base quality even for users with good channel conditions. To optimize QoE, we propose to use multiple multicast sessions with different transmission rates for different SVC layers. The goal is to provide at least the multicast session with acceptable quality to users with bad channel conditions and to provide additional multicast sessions having SVC enhancement layers to users with better channel conditions. The selection of modulation rate for each SVC layer and for each multicast session is achieved with binary integer linear programming depending on network conditions with a goal to maximize global QoE. Results show that our algorithm maximizes global QoE by providing highest quality videos to users with good channel conditions and by guaranteeing at least acceptable QoE for all users.

In this paper, we propose a rate-aware power saving mechanism (RA-PSM) in multi-rate wireless LANs. In RA-PSM, the channel access order is determined depending on the transmission rate (or channel conditions). Since a station with higher transmission rate can request the buffered frames at the access point with higher priority, the overall channel waiting time can be reduced. Preliminary simulation results illustrate that RA-PSM can reduce the average waiting time by 45% compared with the conventional IEEE 802.11 PSM.

In this paper, we analyze the energy efficiency of IEEE 802.11 power save mode in multi-rate environments. We consider two transmission schedules: random and rate-aware schedules. Numerical results demonstrate that the rate-aware schedule can reduce the expected idle time by 36%.

This study presents the design and performance evaluation of a latency bounded and medium access control (MAC) scheme (abbreviated to LB-MAC) for wireless sensor networks that require time-critical communication. LB-MAC provides the predictability for an end-to-sink data delivery time through the time division multiple access (TDMA) multi-channel transmission frame assignment considering the routing path of a tree network. In addition, it uses the traffic-based slot assignment mechanism at every transmission frame to achieve energy efficiency. The feasibility and effectiveness of LB-MAC is demonstrated through extensive simulation based comparisons with prior research.

WiMedia MAC protocol supports fully distributed data communications in high data rate wireless personal area networks (WPANs). WiMedia MAC includes a distributed reservation protocol (DRP) for synchronous traffic and a prioritized contention access (PCA) protocol for asynchronous traffic. Since PCA is based on the carrier-sense multiple access with collision avoidance (CSMA/CA) mechanism, it suffers from low throughput due to collision, especially when the number of devices is large. In this paper, we propose a novel channel access scheme called deterministic channel access (DCA), which determines the transmission order among devices by means of beacon frames. Since all devices follow the deterministic transmission order, collision-free channel access can be achieved and thus the throughput can be significantly improved. Extensive simulation results demonstrate that DCA outperforms PCA in terms of throughput under different situations.

Traffic anomaly detection is one of the most important technologies that should be considered in network security and administration. In this paper, we propose a traffic anomaly detection mechanism that includes traffic monitoring and traffic analysis. We develop an analytical system called WISE-Mon that inspects the traffic behavior by monitoring and analyzing the traffic. We establish a criterion for detecting abnormal traffic by analyzing training set of traffic and applying Fisher linear discriminant method. By using the properties of distributions such as chi-square distribution and normal distribution to the training set, we derive a hyperplane which enables to detect abnormal traffic. Since the trend of traffic can be changed as time passes, the hyperplane has to be updated periodically to reflect the changes. Accordingly, we consider the self-learning algorithm which reflects the trend of traffic and so enables to increase accuracy of detection. The proposed mechanism is reliable for traffic anomaly detection and compatible to real-time detection. For the numerical results, we use a traffic set collected from campus network. It shows that the proposed mechanism is reliable and accurate for detecting the abnormal traffic. Furthermore, it is observed that the proposed mechanism can categorize a set of abnormal traffic into various malicious traffic subsets.