Machine vision has become a revolution in the art criticism of the present time whereby computational tools have been presented to extend, challenge, and redefine long-standing conventions of aesthetic criticism or assessment. The fact that artworks are becoming multimodal and digitized and even hybrid in nature has increased what the algorithms can detect, encode style, and analyze the compositional structures, which has widened the scope of interpretive inquiry. This paper explores the research of using machine vision enabled by neural networks of convolutional convolution, transformers, and vision-language models to model visual information systematically, including texture, color harmony, spatial depth, symbolic and narrative hints. The paper puts AI-driven criticism into a wider framework of seeking the meaning and the role of viewers and machines by applying cognitive theories of perception as well as philosophical approaches to perception and authorship. The methodology consists of the carefully selected collections of paintings, sculptures and digital works of art, as well as the powerful preprocessing pipelines to extract features and perform multimodal embedding. Results show that machine vision improves already existing systems of critique by the quantification of aesthetic qualities, the discovery of latent stylistic similarities, and the multi-layered patterns of interpretation not always evident to human viewers. In addition, AI methods are impacting curatorial practice, providing insights in the form of data to plan an exhibition, research the provenance, and engage the audience. The argument also points out the negative and positive aspects of computational aesthetics, noting the necessity of both human-AI interpretive ecosystems which allow cultural sensitivity but are friendly to people.

The research paper discusses the use of predictive analytics in the planning of sculpture exhibitions to enhance the curatorial decision-making process using predictive decision-making based on the data. The study incorporates the elements of data science, computational aesthetics, theory of art object/virtual display curator, forming a building of modular prediction and using regression, classification, clustering, ensemble learning, and time-series prediction to predict the visitor engagement, create a space layout, and get the sentiment of the audience. There is a high predictive reliability in the system prototype (R2 =0.89, F1 =0.91) that transforms the traditional curating process into a more adaptive and intelligence-driven process. Experiments have discovered that predictive heatmaps, regression graphs, and sentiment trend curves are handy in developing the exhibition into actionable information using complex data. The framework is not only the contributor to spatial performance and visitor satisfaction but also generates a new idea of human-AI collaboration within the creativity of the curators. The findings confirm that predictive analytics can turn the exhibition as an immobile system into a breathing ecosystem that responds to the behavior of the audience and appeal to the emotion, which is another manifestation of a more relevant and substantial solution of the digitalization of museology.

Digital advertising has grown at a very high rate, supporting the necessity of images that are visually appealing and able to attract the consumer on the perception of the image and motivation to purchase a product. The study is a predictive model of visual appeal in advertising photography based on the methods of computational aesthetics, machine learning, and deep learning. The study conceptualizes, in the first instance, visual appeal as a construct of multidimensions that is influenced by composition, lighting, color harmony, subject prominence, emotional tone, and style. A large data set of advertising photos, gathered in several products and media are labeled by a structured labeling protocol, which measures aesthetic quality that is perceived by humans. They use both handcrafted and deep visual descriptors, obtained with the help of pretrained convolutional neural networks and transformer-based encoders, to construct predictive models. Its methodology consists of a preprocessing and normalization pipeline as a whole and two large families of models, CNNs to learn spatial features and transformers to learn contextual features worldwide. Empirical evidence shows that deep representations perform better than handcrafted features at fine-grain aesthetics and that transformer models are more able to predict the associations between the visual complexity and the scores of visual attractiveness. Another limitation noted in the study is the subjectivity of the datasets, cultural biasness, and lack of diversity in advertising situations.

This paper sets out to review the advent of artificial intelligence as an important medium in modern conceptual art practice, with particular reference to its ability to both extend and confuse the traditional operating assumption of ideas being the primary element in conceptual art practice. The study is based on the historical development of conceptualism, starting with the early linguistic and systemic arts and moving to the subsequent computational experimentalism, which orients AI to a tradition of artistic and process-oriented approaches, in which processes, instructions, and networks of meaning take the place of conventional object-based production. The distinctive language, image, and symbolic manipulatory skills of AI present new forms of authorship, autonomy, and indeterminacy and provide artists with the opportunity to create works that predetermine system-directed meaning, algorithmic patterning, and computational aesthetics. By presenting the history of algorithmic practices and the current case study, the paper will show that AI is not only a technical tool but also an active conceptual agent that can act to construct the propositions of art. This incorporates its role as partner, actor, and even proxy author, and leads to a rethinking of the agency of the creative and agency, and purposefulness. The theoretical consequences of the changes throw down challenges to the accepted versions of interpretation, work of art, and the limits of the intelligent in the artistic frames. Finally, the paper concludes that AI has a transformative potential to conceptual art that relates to the possibility of producing novel types of ideas, speculative questions, and bringing immaterial ideas to life.

The paper provides an elaborate, data-driven approach to visual composition analysis of artistic, photographic, and design-based imagery. Conventional compositional ideals, including spatial balance, color harmony, Gestalt grouping, and narrative organization, are converted to quantifiable computational aspects using sophisticated image processing, perceptual modeling, as well as machine learning methods. With the incorporation of the feature-extraction, saliency-analysis, edge-orientation mapping, color-harmony measures, and object-position density modeling, the research findings indicate that the patterns of visual compositions are statistically significant and cross-cut across the genres and across the ages. Clustering and dimensionality reduction of patterns are additional ways of mining latent patterns, stylistic relationships, and trends of evolution hidden within large volumes of data. The framework has high interpretability based on the visual analytics using PCA-based feature embeddings, aggregated saliency heatmaps, line-orientation histograms, radar charts of chromatic attributes, and spatial distribution maps. Results indicate that composition is not intuitively or stylistically defined, but an empirically quantifiable structure, which can be used to improve creative pedagogy, computational aesthetics, and AI-mediated design systems. This piece of art defines a new analytical mode of approach, which crosses the boundary of artistic theory and machine intelligence, providing innovative conclusions on the role of compositional logic in the construction of perception and creative expression.

The current paper includes an original strategy of developing an AI-assisted robotic fabrication of sculptures using multi-material 3D printing, by combining creativity, computation, and automation in the process of modern artwork creation. The paper discusses how we can combine artificial intelligence (AI) with robotic production processes to allow autonomous design production and manipulation of materials. The system is able to increase the artistic freedom by using machine learning algorithms to design and optimize via generative design and optimization and make sure the structures and aesthetics are accurate. An extensive approach was created to combine an AI-based conceptual system with a robotic arm and multi-material print head that was adaptive. The AI model learns on data population of massive information on artistic forms and material behavior, which allows dynamic decision making when fabricating. The suggested pipeline, which includes a digital idea up to a solid sculpture, will comprise real-time sensor input and reinforcement learning to regulate adaptively the print parameters, including layer thickness, deposition speed, and combining materials. Practical work has shown that the system could create complex sculptures of both rigid and flexible materials with new textual and structural variations that could not be produced by traditional sculpture. The study emphasizes how AI-based design intelligence is used to create new forms of human-machine collaboration in the creation of art using robot precision. The findings indicate the possible great impact on the future of the computational aesthetics, digital craftsmanship, and self-fabrication systems, between the creative will and material manifestation.

Combining emotional computers and abstract art results in another perspective to the way people feel upon looking at things that are not pictures. This paper uses machine learning and deep learning to investigate computer methods of determining how abstract art affects individuals. It applies computational aesthetics, and concepts of feeling in the visual perception towards developing a model of the impact of colours, textures, and shapes to the feelings of people. There are several emotion recognition mechanisms namely the RBF-SVM, the random forest, the resnet-50 and the vision transformer, which are tested on a rigorously selected set of abstract artwork to determine how they fare in classifying emotions. Image processing and deep learning techniques are employed to extract features and visual-semantic map which detects emotion indicators in artistic pieces. The method establishes the place of mixed inputs, written, visual, and environmental data to enhance emotional predictions. Data of physiological and psychological feelings are checked to explain whether computer conclusions are similar to the data that people observe. The proposed system design provides an avenue to mood analysis, which includes preparation, all the way up to model evaluation. This will be supported by measures such as accuracy, memory, F1-score, and association with human-answered answers. It tries to relate cognitive psychology and computer modelling by observing the ways in which machines may simulate emotional knowledge in abstract art by analyzing the disparities between algorithmic predictions and human subjectsive judgments.

A computational framework is introduced to map aesthetic categories, highlighting distinctions between human-created competition entries and AI-generated designs. Assessment of architectural aesthetics has traditionally been subjective; however, computational methods now enable the systematic analysis and visualization of design variations. Using 14 aesthetic parameters, the framework applies Principal Component Analysis (PCA) to create visual maps of aesthetic relationships. Two hypotheses are discussed: (1) There is a hegemony of the aesthetic category of the beautiful in mainstream architectural competitions, and (2) that generative models like Stable Diffusion can produce visuals belonging to other aesthetic categories. The computational aesthetic framework is applied to verify both. The first hypothesis is disproven by revealing that human-created designs demonstrate significant aesthetic variation and unpredictability, while the second is confirmed by demonstrating the capacity to produce images in 16 other categories beyond the beautiful. The maps of visual relationships group images by aesthetic categories, encouraging designers to explore and enhance underpopulated areas. The computational aesthetics framework is used to analytically and quantitatively support arguments regarding architectural aesthetics.

In this paper, we document three computational features of the climate model Hector by translating the modalities through which the model predicts climate futures in a game engine. We build upon the computational aesthetics of M. Beatrice Fazi and Matthew Fuller (2016) and suggest additional characteristics of computation to their proposed list. The three features are time series, couplings and cosmograms, which are based on our practical and theoretical inquiry into Hector’s computation and literacy. We develop a framework that employs two processual methods: translating the model’s operations into a game engine and the conceptual as well as transdisciplinary debugging of this transfer across different computational interfaces. Through this framework, we ask: what does the process of translating while debugging Hector reveal about the computational aesthetic of the model, and how can this help inquire into its onto-epistemological imaginary? By thinking and practising through specific features of computational aesthetics, we propose a reimagining of climate computation. We introduce the model, the concept of computational aesthetics, and our research methods to describe the three features of climate computation and their influence on Hector’s onto-epistemological imaginary. Finally, we discuss game engines as a site for critical experimentation with the model and the potential reconfiguration and reimagining of computational aesthetics through what we call “climate engines”.

ABSTRACT This paper examines an artistic paradigm that reinvigorates embodied labor and spontaneous mark-making in human interactions with algorithmic systems. It focuses on the curve as a central method in Sougwen Chung’s collaborative paintings with her five generations of Drawing Operations Units. Using techniques such as supervised learning, computer vision, biofeedback, and full-body tracking, Chung transforms neural activity, physical gestures, and archival drawings into data inputs that guide robotic systems to paint on canvas. Departing from the algorithmic principles of early computer art, Chung’s varying curves improvise fluid forms rather than fixed ones to accompany robotic painting gestures in real-time. The resulting abstract, textured, and spontaneous brushstrokes merge human and machine traces, challenging the assumption that computational aesthetics must appear calculated. Drawing on posthumanism concepts of ‘technosymbiosis’ and ‘cognitive assemblage,’ this paper argues that Chung reconfigures the human-algorithm relationship not as a question of control, but as a dialogic process and embodied communion.

Facial beauty prediction (FBP) is a frontier topic at the intersection of artificial intelligence and computational aesthetics, aiming to enable computers to autonomously predict or assess facial beauty. Currently, while FBP methods have achieved good results on well-processed datasets, they typically exhibit reduced prediction performance on datasets with more unavoidable noisy labels. Diffusion models (DMs) can denoise and reconstruct label encodings, capturing uncertainty in the prediction process through the randomness of their outputs. Therefore, we propose FBP-Diffusion, an improved diffusion model that integrates MobileViT and dynamic loss correction (DLC). Specifically, MobileViT, effective at modeling both detailed and global information, is employed as a conditional information encoder to produce preliminary predictions, which are then fed into the reverse process to guide label generation. DLC is introduced to enhance the model’s denoising capability and robustness, in which the cross-entropy loss is increased by the prediction probabilities of FBP-Diffusion obtained after the reverse process and probability transfer, and then dynamically integrated into the noise estimation loss. Experimental results on four representative facial beauty databases demonstrate that FBP-Diffusion outperforms both conventional DMs and FBP methods, particularly noting a 5.17% accuracy improvement on relatively noisy datasets over state-of-the-art FBP methods.

The research field of computational aesthetics gives crucial contributions to the development of mechanisms for filtering and/or generating value-laden informational content. This paper acknowledges a recognized escalating problem in the development of contemporary informational technologies and presents a practical solution for communicational quality management by employing an innovative approach to the computational aesthetic evaluation (CAE). After discussing the problem and attempted approaches to its alleviation, the paper offers a novel expert solution by presenting an original research approach and its resulting open-sourced model which outperforms its current state-of-the-art competition in semantic and stylistic classification, at the same time providing an idiomatic measure for objective aesthetic evaluation and demonstrating semantically rich and professionally recognized explanatory power which can serve as the solid basis for development of reliable and user friendly content retrieval, generative or auxiliary design applications. Presented model is resource- and privacy-wise utmost conservative. Its use evades all ethical, legal or security concerns that beset all currently prominent models. Its developmental and operational costs are practically nil.

This article examines the impact of artificial intelligence (AI) on the visual production of contemporary art, generating debates about authorship, creativity, and aesthetic legitimacy, while addressing the ethical and labor implications of automated creation. The main objective of this research is to explore the perception, production, and philosophical implications of AI-generated visual art, with an interactive art installation serving as the case study. The study adopted a qualitative and interdisciplinary approach, structured in three phases: a visual and aesthetic analysis of the work using tools from visual studies and computational aesthetics; a critical documentary review on digital art and algorithmic ethics; and an interdisciplinary critical analysis that integrates ethical, cultural, and technological dimensions, considering viewer interaction and the material conditions of production. The results reveal that AI reconfigures visual identity by fragmenting and distorting human images, creating a dynamic perceptual experience influenced by viewer interaction. This process questions traditional notions of authorship, authenticity, and creativity, also highlighting the aesthetic, cultural, and labor implications of contemporary digital art. In conclusion, the research demonstrates that AI not only transforms visual language but also reshapes perceptions of identity and authorship in the digital age, while raising relevant ethical questions about the invisibility of human labor and power structures in algorithm-mediated artistic creation.

The present study offers a stylistic and function-oriented comparison of grief representation in human-authored and AI literature, focusing specifically on Cormac McCarthy’s The Road and Ross Goodwin’s 1 the Road. Based on Roman Jakobson’s functions of language and Alan Turing’s imitation framework, the work seeks to determine whether artificial narratives are capable of imitating emotionally affective mourning or grief expressions with similar functional and poetic depth to those in human literature. Adopting a corpus-based approach, the researchers have applied software tools e.g. spaCy, AntConc, and Voyant Tools for the analysis of affective vocabulary, sentence constructions, poetic modus operandi, and repeated phrase patterns. The analysis shows that, despite a superficial command of poetic techniques (e.g., metaphor, personification, and repetition), features of narrative coherence, psychical continuity, and symbolic characteristic of McCarthy’s human-authored novel are absent from the AI-generated text. The result is an instance of AI grief that is mimetically present, but structurally and emotionally empty, it meets Turing’s imitation requirement but does not express Jakobson’s emotive function. This outcome also points to the inability of AI as it currently stands to produce depth of experience and suggests larger conversations in post human literary theory, authorship, and computational aesthetics. The work provides important perspective on the emerging itineraries of artificial creativity and literary effect.

Digital art analysis is evolving rapidly, with intelligent systems playing a growing role in understanding aesthetic quality and artistic styles. In this work, we present the Hierarchical Multi-Stream Feature Network (HMSFN), a deep learning framework designed to improve the way visual features are extracted and classified across different styles and aesthetic levels. The study is based on a curated dataset of 213,000 digital artworks sourced from online galleries and collections, covering a wide range of creative expressions and thematic categories. To enhance data quality and balance, we applied specialized preprocessing techniques including Contrast-Balanced Normalization, Dominant Color Mapping, and Gradient-Symmetric Scaling. Additionally, Weighted Synthetic Feature Augmentation (WSFA) was introduced to address class imbalance, while an Adaptive Feature Filtering Framework (AFFF) was used to remove redundant features and retain the most informative ones. The model was trained using an 80:20 split and evaluated against several leading deep learning approaches. HMSFN, which combines DenseNet, ConvNeXt, and Vision Transformer in a multi-stream configuration, achieved outstanding results—99.0% accuracy, 98.6% F1-score, 97.5% LCCR, and an AUC of 99.3%. These findings highlight the effectiveness of our approach in capturing complex visual attributes and support its use in digital art classification and computational aesthetics.

This article discusses the emergence of a genre of technologically mediated, computationally networked zoopoetic practices. I approach this discussion through an analysis of contemporary examples of zoopoetry, firstly drawing on print-based examples such as the poetry of e. e. cummings and the poetic animal dialogue of the novelist Laura Jean McKay. I then consider the ways in which digital technologies and digital aesthetics have the potential to add modes and imaginaries to zoopoetic authorial practices. I introduce digital zoopoetics through the creation of two related digital interfaces: The (m)Otherhood of Meep (2023) and The Songbird Speaks (2024-ongoing). These works invite new imaginaries for interspecies signal interpretation through machine learning technology by moving towards generative interspecies translation as its own zoopoetic form. From the practical contribution to zoopoetics that these works make, I offer a non-exhaustive series of suggested affordances of digital and computational aesthetics that come forth as representative of a digital zoopoetry form.

Abstract This study introduces a two-step iterative strategy [40], incorporating the viscosity approximation
technique and h -convexity to refine the algorithm’s precision. Central to analysis is the complex
function A(u) = ur − su2 + bu + sin wt for all u ∈ C, where r ∈ N \ {1}, s ∈ C, and b, w ∈ C \ {0}
with |s| < |w|r−2 and t ∈ [1, ∞). A convergence condition is established in the form of an escape
criterion, enabling the adaptation of the escape-time algorithm to the specific iteration scheme under
consideration. This approach facilitates the generation of intricate fractal structures and analyzes
the impact of parameter variations on their morphology. The refined algorithmic framework enriches
the computational aesthetics of fractals, particularly in the visualization of Julia and Mandelbrot
sets. Furthermore, the relationship between the iterations parameters and two numerical measures,
average number of iterations (ANI) and execution time (ET) in MATLAB, measured in seconds,
offers valuable insights into the influence of these parameters on the graphical representation of
fractal formation.

Latin American digital literature occupies a transformative space in global literary studies, where interactivity, materiality, and computational aesthetics converge with regional cultural and linguistic diversity. However, this field faces critical challenges, including technological obsolescence, linguistic marginalization, and infrastructural precarity. This paper explores these issues through case studies such as Jose Aburto's Paranoico arcoiris and collaborative preservation efforts like the Antología Lit(e)Lat. Drawing on decolonial digital humanities frameworks and theoretical contributions from Salazar Salgado, Gainza, Flores, Hayles, Kozak, and others, the study advocates for inclusive and sustainable preservation practices that reflect the sociotechnical realities of the Global South. This paper highlights its potential to reshape the boundaries of global literary heritage from the personal perspective of the author and his 25 years of work experience in digital and experimental poetry from Perú.

To facilitate comparative analysis of artificial intelligence (AI) and human paintings, we present a unified computational framework combining neural embedding and computational aesthetic features. We first exploit CLIP embedding to provide a projected overview for human and AI painting datasets, and we next leverage computational aesthetic metrics to obtain explainable features of paintings. On that basis, we design a visual analytics system that involves distribution discrepancy measurement for quantifying dataset differences and evolutionary analysis for comparing artists with AI models. Case studies comparing three AI-generated datasets with three human paintings datasets, and analyzing the evolutionary differences between authentic Picasso paintings and AI-generated ones, show the effectiveness of our framework.

Purpose Computational aesthetics involves the application of computer-based methods to analyse and understand various phenomena. Given the significance of university library websites in the advancement of science and knowledge, it is essential to consider their aesthetic qualities in conjunction with their web visibility, particularly regarding search engine optimisation (SEO). This study aims to investigate the relationship between visual complexity as an important aspect of the aesthetics of a website and its SEO rankings of top university library websites. Design/methodology/approach This study employed an analytical survey research design, analysing 35 library websites affiliated with top universities as ranked by Times Higher Education in 2023. Visual complexity was assessed using Python Athec programming, a tool specifically developed for computational aesthetic analysis in social science research. Additionally, the AIOSEO online smart tool was utilised to extract SEO scores. Data analysis was conducted using SPSS and Excel. Findings The study found no significant correlation between visual complexity and SEO rankings, with a significance level of 0.125 indicated by linear regression and correlation analyses. This suggests that top university library websites may not require extensive SEO optimisation due to their established credibility and branding. Notably, even those with lower SEO rankings continue to attract international users. Originality/value This research is distinguished by its innovative use of Python programming to measure user experience in the context of aesthetics. It offers new insights into the field of computational aesthetics for the managers of libraries.