ABSTRACT The classification of tomato ripening stages involves assigning a tomato to a category based on the visual indicators of its maturity. Indeed, the specific number of categories and their attributes are determined by the agricultural standards of each country, which rely on an empirical understanding of visual characteristics. Conversely, automatic unsupervised classification techniques, such as deep learning-based methods, autonomously learn their characteristics. In this research, a comparison is made between expert-based classification and unsupervised classification, with a particular focus on the analysis of the number of clusters and their respective features. Remarkably, this investigation finds an alignment in the number of clusters identified by both methods. This discovery supports the notion that the expert-based classification system is compatible with automated approaches. The outcomes of this research could aid the agricultural sector in refining automatic classification techniques. Furthermore, this work provides the scientific community with valuable insights into the clustering of images by machine learning methods. maturity Indeed country characteristics Conversely techniques learningbased methods expertbased expert features Remarkably approaches Furthermore

RESUMO A classificação do estágio de maturação do tomate envolve atribuir uma classe a um tomate com base nos aspectos visuais da sua maturidade. De fato, o número de classes e suas características são definidos pelas normas do departamento agrícola de cada país, as quais são baseadas no conhecimento empírico sobre as características visuais. Por outro lado, métodos de classificação automática não supervisionados, como aqueles baseados em aprendizado profundo, aprendem suas características de forma independente. Neste trabalho, comparamos a classificação baseada em especialistas com a classificação não supervisionada, analisando particularmente o número de agrupamentos (clusters) e as características de cada um. Surpreendentemente, nosso estudo revela uma coincidência no número de agrupamentos para ambas as abordagens. Com essa descoberta, fornecemos evidências de que a classificação atual baseada em especialistas também é adequada para métodos automáticos. As descobertas deste estudo podem ajudar a indústria agrícola a desenvolver métodos precisos de classificação automática. Além disso, para a comunidade científica, este estudo oferece percepções sobre como os métodos de aprendizado de máquina agrupam imagens. maturidade fato país lado supervisionados profundo independente trabalho supervisionada clusters (clusters Surpreendentemente abordagens descoberta automáticos disso científica imagens

In this paper, we present a new algorithm, namely, a micro artificial immune system (Micro-AIS) based on the Clonal Selection Theory for solving numerical optimization problems. For our study, we consider the algorithm CLONALG, a widely used artificial immune system. During the process of cloning, CLONALG greatly increases the size of its population. We propose a version with reduced population. Our hypothesis is that reducing the number of individuals in a population will decrease the number of evaluations of the objective function, increasing the speed of convergence and reducing the use of data memory. Our proposal uses a population of 5 individuals (antibodies), from which only 15 clones are obtained. In the maturation stage of the clones, two simple and fast mutation operators are used in a nominal convergence that works together with a reinitialization process to preserve the diversity. To validate our algorithm, we use a set of test functions taken from the specialized literature to compare our approach with the standard version of CLONALG. The same method can be applied in many other problems, for example, in text processing.

The flow and distribution of fluids through pores in porous media are governed by their geometry. The self-similar behavior of the structure of these sets has been the subject of numerous studies that have documented the power laws relationship, among the principal measures of pores and solids, and the resolution of the method used for their analysis. A fractal scheme is introduced in order to extract and measure some geometric features of pores, using the mean values of their fractal classifiers, dividing these in global (ftrmagram) and local (reference line) classifiers. The mass fractal dimension (Dm), the spectral dimension orfracton (<img border=0 src="../../../../../img/revistas/rmcg/v23n2/a6s1.jpg">), the Hurst exponent (H) and the lacunarity (Λ) of naturally fractured reservoirs (YNF) of southeastern Mexico, have shown to be statistically different for the three most representative porosity sets: fractures, cavities and mixed porosity. The fractal classifiers, extracted from digital images obtained using X-ray computerized tomography, were useful for porosity classification in different patterns, starting from the core images. Dm and <img border=0 src="../../../../../img/revistas/rmcg/v23n2/a6s1.jpg">must be determined on presegmented images, which distinguish the pore and solid sets, prior to the fractal quantification, while H and Λ can be extracted from the original image, which drastically diminishes the bias in the porosity estimation. All fractal classifiers, including mass dimensions and lacunarity of the porosity patterns specified above for YNF, showed a statistically significant correlation with the porosity of the geological strata as determined by traditional techniques. All our results offer new perspectives for the modeling and forecasting of porosity in the naturally fractured deposits.

El movimiento y la distribución de fluidos a través de los medios porosos están determinados por su geometría. La naturaleza autosimilar de la estructura de estos medios ha sido el objeto de numerosos estudios que han documentado las relaciones de potencia ('power law') entre las principales medidas de poros y sólidos, y la resolución del método utilizado para su análisis. En la presente investigación se introduce un esquema fractal para clasificar los yacimientos naturalmente fracturados (YNF) a partir de imágenes de tomografía computarizada de rayos X. Esta clasificación tiene como propósito extraer y medir algunos rasgos geométricos de los poros tanto a nivel global (firmagrama), como local (líneas de referencia) vía los clasificadores fractales. Los clasificadores fractales, extraídos de las imágenes digitales, fueron útiles para hacer un diagnóstico simple y rápido del tipo de porosidad de un núcleo a partir de su imagen. La dimensión fractal de masa (Dm y <img border=0 src="../../../../../img/revistas/rmcg/v23n2/a6s1.jpg">), la dimensión espectral ofractón (<img border=0 src="../../../../../img/revistas/rmcg/v23n2/a6s1.jpg">),el exponente de Hurst (H) y la lagunaridad (Λ) de los YNF del sureste de México, son estadísticamente diferentes para los tres patrones de porosidad representativos de estos materiales: fracturas, cavidades y porosidad mixta. Para estimar los primeros dos (Dm y <img border=0 src="../../../../../img/revistas/rmcg/v23n2/a6s1.jpg">), es necesario presegmentar la imagen en conjuntos de poros y sólidos, creando una imagen binaria, previo a la cuantificación fractal. En los últimos (H y Λ), la extracción de los parámetros se realiza directamente a partir de las imágenes originales evitando el proceso de segmentación, lo que permite proponer a los clasificadores H y Λ como estimadores más confiables de la porosidad de los YNF. Todos los clasificadores fractales, y en especial la dimensión fractal de masa y la lagunaridad de los tres patrones de porosidad arriba especificados, mostraron una correlación estadísticamente significativa con la porosidad (medida con técnicas tradicionales) de las capas geológicas con distinta capacidad productora de hidrocarburos. Este hecho abre un nuevo panorama para la modelación y pronóstico de la geometría de los YNF.