With the widespread deployment of edge computing, the focus has shifted to machine-centric live video streaming, where endpoint-collected videos are transmitted over networks to edge servers for analysis. Unlike maximizing user's Quality of Experience (QoE), machine-centric video streaming optimizes the machine's Quality of Inference (QoI) by balancing the inference accuracy, inference delay, and transmission latency with video adaptive configuration. Traditional heuristic configuration adaption methods are reliable but unable to respond to erratic network fluctuations. Reinforcement learning (RL) based algorithms exhibit superior flexibility but suffer from exploration mechanisms, resulting in long-tail effects on upload latency. In this paper, we propose FHVAC, which dynamically selects video encoding parameters for live streaming by coherently fusing rule-based and RL-based agent at the feature level. We initially develop a robust rule-based approach for ensuring the low latency in transmission, and employ imitation learning to convert it into a neural network equivalently. Subsequently, we design a novel module to combine the two approaches and assess various fusion mechanisms. Our evaluation of FHVAC across two vision tasks (pose estimation and semantic segmentation) in two scenarios (trace-driven simulation and testbed-based experiment) shows that FHVAC enhances the average QoI, and reduces 10.61%-65.27% latency tail performance compared to prior work.

Crowd counting, a method for measuring crowd sizes, has seen significant advancements with deep learning techniques, which have proven highly effective in accurate estimation. However, the improvement in these methods’ accuracy is frequently achieved at the cost of more intricate model architectures. This paper discusses how to construct high-performance crowd counting models using only simple structures. We propose the Fuss-Free Structure, a simple and efficient architecture with a backbone network and multi-scale feature fusion. It exhibits notable adaptability, ensuring that slight replacing its components do not lead to a substantial decline in performance. The multi-scale feature fusion structure is an uncomplicated design that consists of three distinct pathways, each featuring only a focus transition module. It combines the features from these pathways by directly employing the concatenation operation. By selecting appropriate components, our proposed structure has been trained and evaluated across four public datasets, demonstrating an accuracy that rivals that of existing complex models. Furthermore, a comprehensive evaluation is conducted by replacing the backbones of various models such as CCTrans and the proposed structure with different networks, including MobileNet-v3, ConvNeXt-Tiny, and Swin-Transformer-Small. The experimental results further indicate that excellent crowd counting performance can be achieved with the simple structure proposed by us. Code is available at https://github.com/erdongsanshi/Fuss-Free-structure.

Speech-driven lip synchronization is a crucial technology for generating realistic facial animations, with broad application prospects in virtual reality, education, training, and other fields. However, existing methods still face challenges in generating high-fidelity facial animations, particularly in addressing lip jitter and facial motion instability issues in continuous frame sequences. This study presents VividWav2Lip, an improved speech-driven lip synchronization model. Our model incorporates three key innovations: a cross-attention mechanism for enhanced audio-visual feature fusion, an optimized network structure with Squeeze-and-Excitation (SE) residual blocks, and the integration of the CodeFormer facial restoration network for post-processing. Extensive experiments were conducted on a diverse dataset comprising multiple languages and facial types. Quantitative evaluations demonstrate that VividWav2Lip outperforms the baseline Wav2Lip model by 5% in lip sync accuracy and image generation quality, with even more significant improvements over other mainstream methods. In subjective assessments, 85% of participants perceived VividWav2Lip-generated animations as more realistic compared to those produced by existing techniques. Additional experiments reveal our model’s robust cross-lingual performance, maintaining consistent quality even for languages not included in the training set. This study not only advances the theoretical foundations of audio-driven lip synchronization but also offers a practical solution for high-fidelity, multilingual dynamic face generation, with potential applications spanning virtual assistants, video dubbing, and personalized content creation.

Current face-swapping methods often suffer from issues of detail blurriness and artifacts in generating high-quality images due to the inherent complexity in detail processing and feature mapping. To overcome these challenges, this paper introduces the Amazing Face Transformer (AmazingFT), an advanced face-swapping model built upon Generative Adversarial Networks (GANs) and Transformers. The model is composed of three key modules: the Face Parsing Module, which segments facial regions and generates semantic masks; the Amazing Face Feature Transformation Module (ATM), which leverages Transformers to extract and transform features from both source and target faces; and the Amazing Face Generation Module (AGM), which utilizes GANs to produce high-quality swapped face images. Experimental results demonstrate that AmazingFT outperforms existing state-of-the-art (SOTA) methods, significantly enhancing detail fidelity and occlusion handling, ultimately achieving movie-grade face-swapping results.

In the realm of mobile edge computing, the energy consumption and stability of offloading can be significantly influenced by random task arrivals and constraints imposed by task dependencies. To address this challenge, this study proposes a Multi-stage Generative Adversarial Imitation Learning (Multi-stage GAIL) algorithm. This algorithm engages in multi-stage continuous optimization to handle the offloading and allocation processes of dependent tasks. By integrating a simulation system, the algorithm effectively disentangles dependency relationships and generates a queue of tasks ready for execution. It obtains the optimal decision in multiple stages, effectively addressing the impact of the random task arrivals and dependency constraints on task offloading and allocation. Consequently, it leads to enhancements in system performance compared to the baseline algorithm.