This article proposes a novel approach to object detection in low-illumination conditions, an area where existing methods often fall short. By recognizing the crucial role of image quality in subsequent detection accuracy, we introduce preprocessing steps involving image enhancement and restoration networks. Our detection network then focuses on enhancing both channel and spatial feature information, integrating adaptive attention mechanisms and a Transformer architecture for contextual understanding. A multi-scale fusion strategy is employed to merge these features effectively. Additionally, we implement a partial cross-stage network to facilitate CNN learning while minimizing model size. Experimental results demonstrate the superiority of our approach over previous methods, showcasing significant accuracy improvements, particularly in challenging scenarios. This work underscores the importance of addressing image quality concerns in object detection tasks, especially in low-illumination environments, and offers a promising solution with practical implications across various domains, including autonomous driving and surveillance.

This paper addresses the reflective beamforming design for aerial reconfigurable intelligent surface (ARIS), where the RIS is mounted on an aerial platform at a certain height. Different from the terrestrial counterpart, the ARIS provides additional deployment flexibility but it is susceptible to perturbations such as wind effect and imperfect flight control. Consequently, the reflective beam from the ARIS may not provide the highest signal quality to the intended user. Since random perturbations are not known in prior, we adopt a deterministic uncertainty model where the deviation of the ARIS position from the desired one is bounded in a certain range. Based on this uncertainty model, the reflective beamforming design for maximizing the received signal-to-noise ratio (SNR) is formulated that contains a nonconvex objective function and infinitely many constraints. Two approaches are proposed to find the the phase shifts on the ARIS for constructing robust reflective beams. One resorts to the semi-definite relaxation (SDR) approach and the other leverages the subarray based beam broadening and flattening technique. Simulation results are shown to demonstrate the performance of the proposed robust reflective beamforming.

This article proposes a scalable accelerator for deep learning (DL) implementation on edge computing, which is often limited by power, storage, and computation speed. The accelerator is based on systolic array cores with 126 processing elements (PEs) and optimized for YOLOv3-Tiny with$448\times448$input images. Two multicast (MC) network architectures, feature map multicasting and weight multicasting, are introduced to control data stream distribution within the multicores. Results show that the proposed weight multicast (W-MC) systems outperformed the feature map multicast (FMAP-MC) systems in multicore scenarios, with up to$2.23\times $frame rates per second (FPS). The 4-core W-MC system achieved the best efficiency with an overall frame rate of 13.73 FPS/W and an overall throughput of 35.83 GOPS/W. The 8-core W-MC system delivered the best performance, with a frame rate of 38.50 FPS after normalization to the standard YOLOv3-Tiny network. The proposed accelerator offers better computational efficiency and greater accelerator utilization in real-world inference scenarios, compared to previous state-of-the-art works.

Engineers must consider performance, power consumption, and cost when designing embedded digital systems; furthermore, memory is a key factor in such systems. Code compression is a technique used in embedded systems to reduce the memory usage. BitMask-based code compression is a modified version of dictionary-based code compression. The basic purpose of BitMask is to record mismatched values and their positions to compress a greater number of instructions; it can be used exclusively or incorporated with the reference instructions to decode the codewords. In this paper, we applied a small separated dictionary, and variable mask numbers were used with the BitMask algorithm to reduce the codeword length of high-frequency instructions. In addition, a novel dictionary selection algorithm was proposed to increase the instruction match rates. The fully separated dictionary method was used to improve the performance of the decompression engine without affecting the compression ratio (CR) (the compressed code size divided by original code size). Based on the experimental results, the proposed method can achieve a 7.5% improvement in the CR with nearly no hardware overhead.

We studied the architecture of embedded computing systems from the viewpoint of power consumption in memory systems and used a selective-code-compression (SCC) approach to realize our design. Based on the LZW (Lempel-Ziv-Welch) compression algorithm, we propose a novel cost effective compression and decompression method. The goal of our study was to develop a new SCC approach with an extended decision policy based on the prediction of power consumption. Our decompression method had to be easily implemented in hardware and to collaborate with the embedded processor. The hardware implementation of our decompression engine uses the TSMC 0.18 μm-2p6m model and its cell-based libraries. To calculate power consumption more accurately, we used a static analysis method to estimate the power overhead of the decompression engine. We also used variable sized branch blocks and considered several features of very long instruction word (VLIW) processors for our compression, including the instruction level parallelism (ILP) technique and the scheduling of instructions. Our code-compression methods are not limited to VLIW machines, and can be applied to other kinds of reduced instruction set computer (RISC) architecture.

In this article, we describe a distributed, peer-to-peer gesture recognition system along with a software architecture modeling technique and authority control protocol for ubiquitous cameras. This system performs gesture recognition in real time by combining imagery from multiple cameras without using a central server. We propose a system architecture that uses a network of inexpensive cameras to perform in-network video processing. A methodology for transforming well-designed single-node algorithm to distributed system is also proposed. Applications for ubiquitous cameras can be modeled as the composition of a finite-state machine of the system, functional services, and middleware. A service-oriented software architecture is proposed to dynamically reconfigure services when system state changes. By exchanging data and control messages between neighboring sensors, each node can maintain broader view of the environment with integrated video-processing results. Our prototype system is built on Windows machines, and uses standard video cameras as sensors and local network as a communication channel.

Distributed smart cameras use distributed computing architectures to analyze imagery from physically distributed cameras. Performing real-time distributed analysis of video introduces substantial new challenges, but also provides substantial benefits over server-based approaches. This work describes some of the algorithms and architectures we have developed for tracking using distributed smart camera systems, including fault-tolerance, synchronization, and multi-band fusion.

We propose a new class of methods for VLIW code compression using variable-sized branch blocks with self-generating tables. Code compression traditionally works on fixed-sized blocks with its efficiency limited by their small size. A branch block, a series of instructions between two consecutive possible branch targets, provides larger blocks for code compression. We compare three methods for compressing branch blocks: table-based, Lempel-Ziv-Welch (LZW)-based and selective code compression. Our approaches are fully adaptive and generate the coding table on-the-fly during compression and decompression. When encountering a branch target, the coding table is cleared to ensure correctness. Decompression requires a simple table lookup and updates the coding table when necessary. When decoding sequentially, the table-based method produces 4 bytes per iteration while the LZW-based methods provide 8 bytes peak and 1.82 bytes average decompression bandwidth. Compared to Huffman's 1 byte and variable-to-fixed (V2F)'s 13-bit peak performance, our methods have higher decoding bandwidth and a comparable compression ratio. Parallel decompression could also be applied to our methods, which is more suitable for VLIW architectures.

This paper addresses issues pertaining to the necessity of utilizing formal verification methods in the design of protocols for peer-to-peer multimedia systems. These systems require sophisticated communication protocols, and these protocols require verification. We discuss two sample protocols designed for two distinct peer-to-peer computer vision applications, namely multi-object multi-camera tracking and distributed gesture recognition. We present simulation and verification results for these protocols, obtained by using the SPIN verification tool, and discuss the importance of verifying the protocols used in peer-to-peer multimedia systems

We propose a new variable-sized-block method for VLIW code compression. Code compression traditionally works on fixed-sized blocks and its efficiency is limited by the small block size. Branch blocks-instructions between two consecutive possible branch targets-provide larger blocks for code compression. We propose LZW - based algorithms to compress branch blocks. Our approach is fully adaptive and generates coding table on-the-fly during compression and decompression. When encountering a branch target, the coding table is cleared to ensure correctness. Decompression requires only a simple lookup and update when necessary. Our method provides 8 bytes peak decompression bandwidth and 1.82 bytes in average. Compared to Huffman's 1 byte and V2F's 13-bit peak performance, our methods have higher decoding bandwidth and comparable compression ratio. Parallel decompression could also be applied to our methods, which is more suitable for VLIW architecture.

We describe a peer-to-peer multiple-camera architecture for a distributed real-time gesture recognition system. Previous work attaches multiple cameras to a server This simplifies many design problems but is impractical for real-world installations. Our architecture uses a network of relatively inexpensive cameras to gather images in order to provide high resolution at low cost. Computations are done at the embedded processors in each camera, without using a centralized server. We also propose a methodology for transforming well-defined single-camera algorithms to multiple cameras. We migrate our single-camera gesture recognition system into multiple cameras with slightly overlapped views. In order to minimize the communication bandwidth and power consumption, only selected contours or ellipses information is transmitted between the cameras.