Communication barriers between deaf and hearing individuals can be mitigated through advancements in sign language recognition (SLR) systems. These SLR systems can also improve the user experience of deaf people when interacting with conversational systems that could emerge in the near future. This work explores a landmark-based approach for word classification within an SLR system. The study investigates the impact of a novel data-cleaning methodology on model performance during training. Specifically, a data cleaning process focused on video trimming and sign placement correction is shown to significantly improve dataset quality, resulting in more accurate classification. This cleaner data not only facilitated a more stable training process for the RNN model but also effectively delayed the onset of overfitting compared to a model trained on the original data. The findings highlight the critical role of data quality, particularly when dealing with the limitations inherent to small datasets commonly encountered in SLR tasks. The contribution of this study lies in demonstrating how targeted data cleaning enhances model stability and performance in resource-limited SLR systems.

Thispaper compares the usability of various Apple MacBook Pro laptops were tested for basic machine learning research applications, including text-based, vision-based, and tabular data. Four tests/benchmarks were conducted using four different MacBook Pro models—M1, M1 Pro, M2, and M2 Pro. A script written in Swift was used to train and evaluate four machine learning models using the Create ML framework, and the process was repeated three times. The script also measured performance metrics, including time results. The results were presented in tables, allowing for a comparison of the performance of each device and the impact of their hardware architectures.

In this paper, the authors have compared all of the currently available Apple MacBook Pro laptops, in terms of their usability for basic machine learning research applications (text-based, vision-based, tabular). The paper presents four tests/benchmarks, comparing four Apple Macbook Pro laptop versions: Intel based (i5) and three Apple based (M1, M1 Pro and M1 Max). A script in the Swift programming language was prepared, whose goal was to conduct the training and evaluation process for four machine learning (ML) models. It used the Create ML framework—Apple’s solution dedicated to ML model creation on macOS devices. The training and evaluation processes were performed three times. While running, the script performed measurements of their performance, including the time results. The results were compared with each other in tables, which allowed to compare and discuss the performance of individual devices and the benefits of the specificity of their hardware architectures.

Massive image datasets are often required for the proper functioning of Machine Learning (ML) and Computer Vision (CV) applications. This paper offers a solution to computational challenges in the Image Processing of satellite imagery, by proposing an optimization procedure. The presented approach is verified by an exemplary Python implementation, constituting a standalone tool for automating the dataset creation and labeling, including the extraction of road network data from the national satellite cartography provider. The collected data include detailed road maps along with the parcel information obtained via WebMapService endpoints. The method presented in this paper involves three basic steps: road segmentation (using the Shapely module) to facilitate handling high-resolution orthoimagery, and then a modified Region-of-Interest approach, i.e., removing irrelevant areas, with only roads remaining. This results in obtaining file sizes that are significantly smaller. The presented algorithm also involves asynchronous tile downloading, which, combined with the masking of irrelevant areas, improves not only the efficiency but surprisingly also the accuracy of subsequent ML/CV procedures. The research results of the paper reveal substantial file size reduction, and improved processing efficiency, thus making the optimized geospatial graphical data more practical for ML/CV applications, while still maintaining the original data quality and relevance of the analyzed parcels or infrastructure.

This study assesses roadside grass biomass harvesting and allocation to collection centers using spatially explicit methods for preliminary district-heating supply-chain planning. Road-network data combined with mowing data for county roads were used to estimate theoretical, technically recoverable, and annual energy potentials in Kołobrzeg County, Poland, while high-resolution aerial imagery was used to identify roadside sections where biomass cannot be harvested, such as bridges, paved shoulders, gravel surfaces, and bicycle paths. Dijkstra’s algorithm calculated shortest network distances, after which a custom assignment procedure allocated road nodes and associated biomass to collection centers using either the baseline distance-based rule or its capacity-constrained extension. The estimated annual energy potential ranged from 14.9 to 38.1 TJ/a among the analyzed municipalities. In the two-center scenarios, the number of nodes assigned to the southern center increased from 25 under its $3000\mathrm{t}$ limit to 41 under equal $5875\mathrm{t}$ limits at both centers. The results show that combining aerial-image interpretation, biomass-to-energy estimation, and network-based allocation provides a transparent framework for identifying priority collection areas and organizing roadside biomass transport. The proposed workflow can support local authorities in preliminary biomass supply-chain planning and can be adapted to other regions with suitable spatial data.