Automatic recognition of features in digital images has become a central topic in the field of cultural heritage diagnostics.AI-based models are being increasingly applied to the analysis of infrared reflectography and thermographic data.They show great promise in automating time-consuming manual analyses and improving the objectivity and repeatability of diagnostic assessments.This work proposes 4 specialized classifiers for nails and detachments in work of arts.In-situ active thermography measurements are used for training proposed models.AI models for nail classification reached accuracy of 96.03 % and 93.65 % using planar composite thermal images and volumetric raw data as inputs, respectively.AI models for detachment classification reached accuracy of 87 % and 57 % using planar composite thermal images and volumetric raw data as inputs, respectively.This work was supported by the Slovak Scientific Grand Agency VEGA under the contract 2/0135/23 "Intelligent sensor systems and data processing".Funded by the EU NextGenerationEU through

Real-time strategy games are popular in AI research and education.Among them, Starcraft: Brood War (SCBW) is particularly well known.Recently, the largest known SCBW game replay dataset STARDATA was published.We classify player strategies used in the dataset for all 3 playable races and all 6 match-ups.We focus on early to mid-game strategies in matches less than 15 minutes long.By mapping the classified strategies to replay files, we label the files of the dataset and make the labeled dataset available.

Real-time strategy games are currently very popular as a testbed for AI research and education.StarCraft: Brood War (SC:BW) is one of such games.Recently, a new large, unlabeled human versus human SC:BW game replay dataset called STARDATA was published.This paper aims to prove that the player strategy diversity requirement of the dataset is met, i.e., that the diversity of player strategies in STARDATA replays is of sufficient quality.To this end, we built a competitive SC:BW agent from scratch and trained its strategic decision making process on STARDATA.The results show that in the current state of the competitive environment the agent is capable of keeping a stable rating and a decent win rate over a longer period of time.It also performs better than our other, simple rule-based agent.Therefore, we conclude that the strategy diversity requirement of STARDATA is met.

Segmentation is the key computer vision task in modern medicine applications.Instance segmentation became the prevalent way to improve segmentation performance in recent years.This work proposes a novel way to design an instance segmentation model that combines 3 semantic segmentation models dedicated for foreground, boundary and centroid predictions.It contains no detector so it is orthogonal to a standard instance segmentation design and can be used to improve the performance of a standard design.The presented custom designed model is verified on the Gland Segmentation in Colon Histology Images dataset.

New high throughput floating-point dividers implemented in FPGA based on different fast computation division algorithms are proposed. The hardware implementations uses 32-bit floating-point single precision. The implementations include both multiplicative inverse and division. The proposed hardware implementations are designed with high computation speed and throughput. They are oriented for high computation demanding applications with multiple division computations in short sequences.

New dedicated hardware architecture for object tracking preprocessing optimized and implemented in FPGA is proposed. It calculates the background image and dual form foreground image while reducing noise in the process. Used algorithms are optimized, the number of mathematical operations are reduced and multiplications are eliminated. 1280 × 1024 pixels optimized multiplier less hardware implementation is composed of 5 small dedicated architectures inter connected by multiplexers and internal registers. The proposed architecture is easily scalable. It is oriented for security tracking applications working in outdoor environment; however, it can be used in any image processing applications as a visual data preprocessing stage. It is resolution and frame rate independent and suitable for all high resolution and multiple camera systems. Optimization for FPGA makes it also suitable for reconfigurable computing and reconfigurable systems.

MDCT is the basic processing component for high quality audio compression. It is also the most computationally intensive operation in the vast majority of audio compression standards. Mostly used audio standard for audio compression is still MP3. This paper presents new implementations of five MDCT / IMDCT architectures with different parallelization levels for MP3. Implementation utilize UMC 90 nm CMOS technology. Design was optimized for low power applications. Low power libraries and clock gating technique were used for power reduction. All IP cores are capable of computing forward and backward MDCT and this feature makes them universal in multimedia SoCs for accelerating the MP3 audio compression/decompression.

Compression of digital audio signals has become increasingly more important with the advent of fast and inexpensive microprocessors and digital signal processors. Several compression schemes were developed and well established. Most of them adopt MDCT/IMDCT. This paper presents a new MDCT IP core generator. The software tool generates several MDCT architectures with adjustable parameters for FPGA-based design as well as computation precision and area estimations