What does this research mean for the field?

The CNN-VISCOM model achieves 96% accuracy in classifying modern art design elements, significantly outperforming existing methods. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

The research aims to develop an effective computational model for interpreting complex visual data in modern art designs.

March 16, 2026Open Access

An intelligent model for visual communication in modern art

Key Points

The research aims to develop an effective computational model for interpreting complex visual data in modern art designs.
Development of a hybrid CNN-STING framework named CNN-VISCOM.
Utilization of STING for multi-resolution grid-based clustering of visual datasets.
Preprocessing includes grid partitioning, image normalization, and noise filtering.
Benchmarking against baseline methods using multiple performance metrics.
CNN-VISCOM achieves 96% accuracy, outperforming LSTM (86%) and CNN-RNN (85%).
Precision rate of 95%, recall of 94%, and F1-score of 95.5% confirm superior classification.
The hybrid model enhances analytical reliability and scalability in design interpretation.

Abstract

The increasing complexity of modern art design challenges traditional visual communication systems. As contemporary Visual Communication artworks integrate multilayered graphics, typography, dynamic color structures, and spatial compositions, existing manual and machine-learning approaches struggle to efficiently interpret large-scale visual data. This creates a growing need for intelligent computational models capable of extracting meaningful visual patterns while managing the massive volume, variability, and ambiguity inherent in modern art design. To address this challenge, this study proposes a hybrid CNN–STING framework, termed CNN-VISCOM, which combines the Statistical Information Grid clustering technique with convolutional neural networks. The STING algorithm reduces computational complexity by performing multi-resolution grid-based statistical clustering, allowing the system to organize large visual datasets into meaningful, low-dimensional representations. The clustered output is then processed by a CNN architecture that extracts spatial, color, texture, and semantic design features, enabling precise classification across various Visual Communication design categories. The dataset consists of several thousand modern art design samples are including posters, logos, packaging layouts, text-based designs, and scene illustrations collected from verified digital-art repositories and Visual Communication design archives. Preprocessing includes STING-based grid partitioning, image normalization, resizing, and noise filtering to ensure structural consistency. The proposed model is benchmarked against existing baselines methods using accuracy, precision, recall, F1-score, and statistically validated using a one-way ANOVA test. The results show that CNN-VISCOM achieves 96% accuracy, outperforming LSTM (86%), CNN-RNN (85%), and DCGAN (78%). Precision (95%), recall (94%), and F1-score (95.5%) further confirm its superior ability to recognize and classify Visual Communication design elements across multiple styles and categories. Overall, the hybrid suggested approach significantly enhances analytical reliability, scalability, and interpretability in Visual Communication modern art design. This impact establishes the framework as a robust foundation for future AI-driven design systems capable of supporting intelligent creation, automated analysis, and advanced design decision-making.

Bookmark

View Full Paper

Cite This Study

Haiyan Zhang (Sat,) studied this question.

synapsesocial.com/papers/69b79ea18166e15b153ac373 https://doi.org/https://doi.org/10.1007/s44163-026-00937-6

Bookmark

View Full Paper