This paper presents an improved YOLOv12 model that integrates the Speckle Attention Mechanism (SAM) to enhance small target detection in complex scenes. The model utilizes a spatial-channel collaborative attention module to generate energy weight maps, achieving a lightweight design with a channel compression ratio of 1/8 and only a 1.2 % increase in parameters. Experiments on a ...

The contemporary business landscape is undergoing a profound transformation, driven by the relentless tide of digitalization. This paper explores the intricate interplay between digital transformation, evolving management practices, and the catalytic role of Artificial Intelligence (AI). It argues that AI is not merely a technological tool but a foundational force that is redefining ...

This paper presents a hardware implementation of the Column-Line Code, an error correction code. The design incorporates three key modules: the encoder, the decoder, and an error injection model for simulation and verification.

In recent years, Near Field Communication (NFC) technology has been widely applied in the field of modern payment systems. However, in practical applications, NFC technology is vulnerable to attacks such as eavesdropping, data tampering, data corruption, cloning, and phishing, which can lead to the leakage of users' private data and pose serious threats to financial information ...

Performance Time Series (PTS) are widely used to monitor system health, yet their analytical value is often limited by a lack of structured event records. This work argues that PTS analyzability is not an inherent property, but a progressive, human-curated and context-dependent property determined by the presence and richness of causal event annotations. We propose a three-level ...

The integration of Artificial Intelligence (AI) into business operations is no longer a mere technological upgrade but a profound cultural catalyst. This paper examines the bidirectional relationship between AI and corporate culture, arguing that while a supportive culture is a prerequisite for successful AI adoption, the technology itself simultaneously acts as a powerful force ...

The possibilities of artificial intelligence methods, in particular multilayer neural networks, are being studied for object recognition in X-ray images obtained at low photon fluxes. The achieved accuracy level (81–90 %) allows such approaches to be considered promising for further integration into intelligent data processing systems for medical and industrial radiography.

For resource-constrained platforms, this paper presents an 8 × 8 blockbased parallel implementation of the DCT/IDCT. The method decomposes the 2D transform into two 1D transforms, uses a slice-safe preallocated 3D buffer together with block-level parfor parallelism, and avoids contention and copying overhead from temporary arrays. Using a batch of 10 images as an example, the ...

This paper introduces a novel side-channel attack named the M-Listening Mouse, which exploits the high precision and sensitivity of modern optical mouse sensors to secretly eavesdrop on user speech by detecting minute vibrations on the desktop surface. Audio signals can cause subtle vibrations on the desktop, which can be sensed by the optical sensors of the mouse. Attackers can ...

A bioinspired ACO algorithm for routing in multi-agent IoT networks is proposed. Route selection uses pheromone concentration, node distances, and residual energy. Simulations show that this approach yields shorter routes and balanced energy usage, improving network robustness and lifetime.