Controller Area Network (CAN) is the de facto standard for in-vehicle networks. However, it is inherently vulnerable to various attacks due to the lack of security features. Intrusion detection systems (IDSs) are considered effective approaches to protect in-vehicle networks. IDSs based on advanced deep learning algorithms have been proposed to achieve higher detection accuracy. However, those systems generally involve high latency, require considerable memory space, and often result in high energy consumption. To accelerate intrusion detection and also reduce memory and energy costs, we propose a new IDS system using Binarized Neural Network (BNN). Compared to full-precision counterparts, BNNs can offer faster detection, smaller memory cost, and lower energy consumption. Moreover, BNNs can be further accelerated by leveraging Field-Programmable Grid Arrays (FPGAs) since BNNs cut down the hardware consumption. The proposed IDS is based on a BNN model that suits CAN traffic messages and takes advantage of sequential features of messages rather than each individual message. We also explore various design choices of BNN, including increasing network width and depth, to improve accuracy as BNNs typically sacrifice accuracy. The performance of our IDS is evaluated with four different real vehicle datasets. Experimental results show that the proposed IDS reduces the detection latency (3 times faster) on the same CPU platform while maintaining acceptable detection rates compared with full-precision models. We also examine the proposed IDS on multiple platforms, and our results show that using FPGA hardware reduces the detection latency dramatically (128 times faster) with lower power consumption compared to an embedded CPU device. Furthermore, we evaluate BNNs with different designs. Our experimental results demonstrate that wider or deeper models definitely improve accuracy at the cost of increased latency and model sizes to varying degrees. Applications are recommended to choose the appropriate model design they need depending on available resources they have.

Human pose estimation (HPE) is a crucial computer vision task with a wide range of applications in sports medicine, healthcare, virtual reality, and human-computer interaction. The demand for real-time HPE solutions necessitates the development of efficient deep-learning models that can be deployed on resource-constrained devices. While a few surveys exist in this area, none delve deeply into the critical intersection of efficiency and performance. This survey reviews the state-of-the-art efficient deep learning approaches for real-time HPE, focusing on strategies for improving efficiency without compromising accuracy. We discuss popular backbone networks for HPE, model compression techniques, network pruning and quantization, knowledge distillation, and neural architecture search methods. Furthermore, we critically analyze the existing works, highlighting their strengths, weaknesses, and applicability to different scenarios. We also present an overview of the evaluation datasets, metrics, and design for efficient HPE. Finally, we identify research gaps and challenges in the field, providing insights and recommendations for future research directions in developing efficient and scalable HPE solutions.

Traditional limb kinematic analysis relies on manual goniometer measurements. With computer vision advancements, integrating RGB cameras can minimize manual labor. Although deep learning-based cameras aim to offer the same ease as manual goniometers, previous approaches have prioritized accuracy over efficiency and cost on PC-based devices. Nevertheless, healthcare providers require a high-performance, low-cost, camera-based tool for assessing upper and lower limb range of motion (ROM). To address this, we propose a lightweight, fast, deep learning model to estimate a human pose and utilize predicted joints for limb ROM measurement. Furthermore, the proposed model is optimized for deployment on resource-constrained edge devices, balancing accuracy and the benefits of edge computing like cost-effectiveness and localized data processing. Our model uses a compact neural network architecture with 8-bit quantized parameters for enhanced memory efficiency and reduced latency. Evaluated on various upper and lower limb tasks, it runs 4.1 times faster and is 15.5 times smaller than a state-of-the-art model, achieving satisfactory ROM measurement accuracy and agreement with a goniometer. We also conduct an experiment on a Raspberry Pi, illustrating that the method can maintain accuracy while reducing equipment and energy costs. This result indicates the potential for deployment on other edge devices and provides the flexibility to adapt to various hardware environments, depending on diverse needs and resources.