Ai-Assisted Student Evaluation in Visual Art Programs

The paper introduces an AI-based system that helps students in visual art education to be evaluated with the help of computational intelligence and pedagogical evaluation to achieve a better degree of objectivity, inclusivity, and creative insight. Conventional methods of art evaluation can tend to be subjective in nature resulting in inconsistency in grading and variation in feedback. The offered system presents a multimodal evaluation pipeline, that is, visual, structural, and stylistic parts of student art are analyzed with the help of convolutional neural networks (CNNs), transformer-based models, and aesthetic perception algorithms. Model training and validation are performed using a training dataset that includes student artworks, expert rubrics, and process logs. The AI model develops multi-criteria scores in terms of creativity, technique, aesthetic quality, and originality dimensions and guarantees the correspondence to the standards of education and outcome-based learning goals. A feedback generation component translates the outputs of the model to have pedagogically significant results, which is beneficial to learners and instructors. The focus is made on the transparency, explainability, and bias mitigation to make sure that the evaluative process of the AI can support but not restrict the artistic freedom

Convolutional neural network · Originality · Perception · Aesthetic Perception and Analysis · Art Education and Development · Digital Media and Visual Art

• Artificial intelligence-supported art education

  Min‐Chi Chiu, Min-Chi Chiu et al.•Interactive Learning Environments•2024

  Add to list