neural network

Synthesis of Multi-Bit Pyramidal Adders on FPGA

The paper analyses the system characteristics and functional capabilities of multi-bit pyramidal adders that can be used in the structures of discrete perceptron of modern neural networks for summing weight coefficients and input signals. The methodology for designing multi-bit adders using flow and spatial-temporal graphs is described. Algorithmic and recursive pyramidal adders have been developed and their main system characteristics have been determined.

Hardware Optimization of Video Quality Improvement Methods Based on Deep Neural Networks

The paper addresses various aspects of optimizing deep video enhancement models for efficient execution on modern hardware. The focus is on a multi-frame generative network with multi-scale structure and frame-by-frame smoothing (MST-GAN). A comprehensive hardware acceleration strategy is proposed, which includes structural thinning, quantization (FP16/INT8), pipeline, parallelization, and model compilation using TensorRT. A comparative analysis is performed before and after optimizations, including changes in FPS, latency, memory consumption, and FLOPs.

IMAGE QUALITY ENHANCEMENT USING NEURAL NETWORKS FOR VARIOUS TYPES OF DISTORTION

The paper considers the problem of enhancing image quality using neural networks with different types of multi-level distortions - contrast degradation, adding noise of various natures, image compression, etc. The widespread TID2013 database, which contains both original images and images modified using various types of distortions (25 basic images, 24 types of distortions, and 5 of their levels), was used as the image database for training neural networks. This database was divided into training (480 images), validation (360), and test (120) images.

Improving the localization of mobile robot by filtering dynamic objects using camera image segmentation

This paper presents an approach to improving mobile robot localization by filtering dynamic objects using camera image segmentation. The proposed algorithm integrates a Particle Filter with state-of-the-art computer vision techniques, specifically employing the YOLO model for segmentation, which effectively differentiates static elements of the environment from moving objects. This approach reduces the impact of noisy data and enhances localization accuracy in dynamic conditions, which is crucial for the reliable autonomous operation of mobile robots.

Building and Optimizing Lightweight Generative Adversarial Neural Networks to Enhance Video Quality in the Client Devices Using Webgpu

The paper considers problems for the tasks of improving the quality of digital video images for cloud environments, as well as on the client side using generative adversarial neural networks (GANs) adapted for work in the browser. A method is proposed that uses WebGPU for accelerated execution of convolutional calculations, which allows to increase the resolution and improve the quality of low-quality video in real time without significant load on servers.

APPLICATION OF AN ADAPTIVE NEURAL NETWORK FOR THE IDENTIFICATION OF FRACTIONAL PARAMETERS OF HEAT AND MOISTURE TRANSFER PROCESSES IN FRACTAL MEDIA

Physics-Informed Neural Networks (PINN) represent a powerful approach in machine learning that enables the solution of forward, inverse, and parameter identification problems related to models governed by fractional differential equations. This is achieved by incorporating residuals of operator equations, boundary, and initial conditions into the objective function during training.

CREATING A FACIAL RECOGNITION SYSTEM BASED ON A NEURAL NETWORK THAT WORKS IN REAL TIME

The article is devoted to the theoretical foundations and practical aspects of facial recognition using neural networks. The paper analyzes various approaches to feature extraction from facial images, as well as methods for training neural networks for recognition. Special attention is paid to problems related to variations in lighting, facial expressions, and other factors affecting recog- nition accuracy. The paper describes the creation and optimization of a neural network for real-time human face recognition.

PERFORMANCE ANALYSIS OF CNN-ENHANCED GENETIC ALGORITHM FOR TOPOLOGICAL OPTIMIZATION IN METAMATERIAL DESIGN

The Combination of Convolutional Neural Networks (CNN) and Genetic Algorithms (GA) provides a promising approach for topological optimization of complex lattice structures. Lattice structures are commonly used as base in the design of high-performance metamaterials. This paper presents a review of the effectiveness and efficiency of the CNN-GA method. We will examine the ability of the method to generate optimal complex structures while minimizing material usage. CNN is utilized mainly as an analysis instrument.

Comparative Analysis of Maximum Power Point Tracking Algorithms for Photovoltaic Panels

The growing demand for electricity and the need for environmentally friendly energy sources are driving the active development of renewable technologies, with solar energy playing a leading role. Photovoltaic (PV) systems are capable of converting solar radiation into electrical energy; however, their efficiency depends on the ability to adapt to changing external conditions, such as solar irradiance and ambient temperature.

Optimization of the Algorithm Flow Graph Width in Neural Networks to Reduce the Use of Processor Elements on Single-board Computers

The article presents a method for optimizing the algorithm flow graph of a deep neural network to reduce the number of processor elements (PE) required for executing the algorithm on single-board computers. The proposed approach is based on the use of a structural matrix to optimize the neural network architecture without loss of performance. The research demonstrated that by reducing the width of the graph, the number of processor elements was reduced from 3 to 2, while maintaining network performance at 75% efficiency.