This paper examines and compares two neural networks, U-Net-Attention and SegGPT, which use different attention mechanisms to find relationships between different parts of the input and output data. The U-Net-Attention architecture is a dual-layer attention U-Net neural network, an efficient neural network for image segmentation. It has an encoder and decoder, combined connections between layers and connections that pass through hidden layers, which allows information about the local properties of feature maps to be conveyed. To improve the quality of segmentation, the original U-Net architecture includes an attention layer, which helps to enhance the search for the image features we need. The SegGPT model is based on the Visual Transformers architecture and also uses an attention mechanism. Both models focus attention on important aspects of a problem and can be effective in solving a variety of problems. In this work, we compared their work on segmenting cracks in road surface images to further classify the condition of the road surface as a whole. An analysis and conclusions are also made about the possibilities of using architectural transformers to solve a wide range of problems.

Keywords: machine learning, Transformer neural networks, U-Net-Attention, SegGPT, roadway condition analysis, computer vision

This article describes the first stage of the research work on the development of an FPGA-based camera for vehicle identification tasks, which are widely used in automated weight and size control points. Since the FPGA is an alternative to conventional processors, which features the ability to perform multiple tasks in parallel, an FPGA-equipped camera will be able to perform the functions of detecting and identifying vehicles at the same time.Thus, the camera will not only transmit the image, but also transmit the result of processing for problem-oriented control systems, decision-making and optimization of data flow processing, after which the server will only need to confirm or deny the results of the camera, which will significantly reduce the image processing time from all automated points of weight and size control.In the course of development, a simple VGA port board, a static image program for displaying it on a monitor in 640x480 resolution, and a pixel counter program were implemented. EP4CE6E22C8 is used as FPGA, the power of which is more than enough to achieve the result.

Keywords: system analysis methods, optimization, FPGA, VGA adapter, Verilog, recognition camera, board design, information processing, statistics