Mobile video calls play an essential role in our daily lives. However, in bandwidth-limited scenarios (e.g., inadequate cellular coverage, congested satellite links, and metered connections), users often experience poor quality of experience (QoE) during video calls. While recent advances in deep learning have demonstrated significant improvements in video compression over traditional methods, existing approaches are ill-suited for bidirectional video streaming on smartphones. The primary challenge lies in simultaneously achieving high video quality, computational and bandwidth efficiency, and practical usability on constrained mobile devices. In this work, we present SlimFit-Gens, the first practical video calling system for smartphones capable of delivering real-time 480p video at as low as 30 kbps. SlimFit-Gens addresses the challenge with joint algorithm and system-level optimizations. The core technique is a fine-grained model personalization design tailored for mobile video calling, enabling high-fidelity video generation at low model complexity. SlimFit-Gens achieves effective personalized adaptation through a novel two-stage personalization mechanism working upon an optimized model architecture. It also incorporates a privacy preserving, resource-efficient system design, featuring TEE-based (e.g., Confidential VM/NVIDIA Confidential Computing) fine-tuning on the server side and heterogeneity-aware inference on the device side. We implement SlimFit-Gens on four commercial off-the-shelf (COTS) smartphones with different system-on-chip (SoC) configurations and conduct extensive evaluations. Compared to prior work, SlimFit-Gens simultaneously improves generation quality with a 0.09-0.12 reduction in LPIPS and system efficiency through a 1.6-1.8× increase in video frame rate.

Commonly used semi-supervised medical segmentation networks usually use consistent learning under different data perturbations to regularise training, ignoring the multiscale information of the data itself. Therefore, this paper proposes a new network based on coarse and fine dual training streams(CF-UNet), which consists of a backbone network and auxiliary learning branches(ALB). Our approach has the following two novel designs: 1) We design a simple and effective coarse- fine dual training streams. Specifically, in the coarse training stream, we improve the robustness and generalisation of the model by establishing regularisation between different strong and weak perturbation views. In the fine training stream, we introduce an auxiliary learning branch to improve the prediction performance of the backbone network.2) In the ALB module, we design the channel spatial fusion attention module (CSMA) and multiscale large kernel convolutional attention (MS-LKA) to perform feature extraction and fusion from a variety of scales. We evaluate our proposed method on ACDC and DRIVE datasets and numerous experiments have shown that our CF-UNet outperforms state-of-the-art networks. Code is available at https://github.com/slz-bit/CF-UNet.

Federated learning (FL), as a distributed training paradigm, has drawn great attention from both academia and industry. Recently, privacy and security concerns have been raised for FL. Despite many efforts to protect privacy and security, an FL framework that can systematically provide privacy and security guarantees is lacking. In this work, we present TTFL, a trustworthy FL framework in practice to defend the security and privacy issues based on Arm Confidential Compute Architecture (CCA). TTFL has two core designs. (1) It achieves a high-availability privacy protection based on flexible Trusted Execution Environments (TEEs). It leverages the resource-rich and conveniently accessed features of the latest TEE on Arm CCA, combined with our TEE secure interconnection design, to enable the whole FL process performed in distributed TEEs, which efficiently protects parameter confidentiality and protocol integrity. (2) It achieves effective security protection by proposing an effective poisoning-resisted secure aggregation scheme and protecting it within TEE. The new proposed secure aggregation combines the advantages of existing defenses and is placed in the flexible TEE to ensure a secure, effective, and non-bypassable aggregation procedure. We implement a prototype of TTFL and evaluate it regarding security, privacy, and system performance. Evaluation results show that TTFL can comprehensively and efficiently address the main privacy and security threats in FL. For instance, compared with previous work, it improves the model accuracy by 1.9% and reduces the attack success rate by 79.7% on the CIFAR-10 dataset with only about 19.8% training time overhead.

ARM TrustZone is widely deployed on commercial-off-the-shelf mobile devices for secure execution. However, many Apps cannot enjoy this feature because it brings many constraints to App developers. Previous works have been proposed to build a secure execution environment for developers on top of TrustZone. Unfortunately, these works are still not a fully-fledged solution for mobile Apps, especially for the emerging intelligent Apps. To this end, we propose LEAP, which is a lightweight developer-friendly TEE solution for mobile Apps. LEAP enables isolated codes to execute in parallel and access peripheral (e.g., mobile GPUs) with ease, flexibly manages system resources upon different workloads, and offers the auto DevOps tool to help developers prepare the codes running on it. We implement the LEAP prototype on the off-the-shelf ARM platform and conduct extensive experiments on it. The experimental results show that Apps can be adapted to run with LEAP easily and efficiently. Compared to the state-of-the-art work along this research line, LEAP can achieve an average 3.57× speedup in supporting intelligent Apps using mobile GPU acceleration.

While ultrasound- and magnetic resonance-based elastography techniques have proved to be powerful biomedical imaging tools, most approaches assume isotropic material properties. In this paper, a general framework is developed for tomography-based anisotropic elastography. An anatomically well-motivated piece-wise homogeneous model is proposed to represent a class of biological objects consisting of different regions. With established tomography modality, static displacements are measured on the entire external and internal boundaries, and the force distribution is recorded on part of the external surface. A principle is proposed to identify the anisotropic elastic moduli of the constituent regions with the obtained boundary measurements. The reconstruction procedure is optimization-based with minimizing an objective function that measures the difference between the predicted and observed displacements. Analytic gradients of the objective function with respect to the elastic moduli are calculated using an adjoint method, and are utilized to significantly improve the numerical efficiency. Simulations are performed to identify the elastic moduli in a breast phantom consisting of soft tissue and a hard tumor. For isotropic phantom, one set of the boundary measurements enables unique reconstruction results for the tissue and tumor. For anisotropic phantom, however, multiple sets of the measurements corresponding to different deformation modes become necessary.