Print, check, repeat: digital quality by design for 3D-printed medicines using OpenAI models

Area of Science:

• Additive Manufacturing
• Materials Science
• Artificial Intelligence

Background:

• Semi-solid extrusion three-dimensional printing (SSE-3DP) is a versatile technology with broad industrial applications.
• Standardized quality control (QC) strategies are lacking, hindering routine implementation of SSE-3DP.
• Post-processing inspection is difficult due to variations in material, geometry, and optical properties.

Purpose of the Study:

• To develop a systematic and scalable QC framework for SSE-printed patches.
• To integrate colorimetric analysis with AI-assisted machine vision for enhanced quality assessment.
• To improve the reproducibility and automation of QC in SSE-3DP.

Main Methods:

• Utilized three hydrogel ink classes (starch, pectin, gelatin) for printing model patches.
• Quantified color contrast using CIEDE2000 metric and validated with image analysis.
• Employed AI tools (ChatGPT-5, Perplexity Pro) for color prediction, image segmentation, and geometric comparison (IoU).

Main Results:

• Optimized color combinations improved edge detection, segmentation accuracy, and defect identification.
• AI-assisted workflow accurately quantified infill accuracy and structural deviations without specialized hardware.
• Successfully differentiated between well-printed and defective patches using the developed QC strategy.

Conclusions:

• The developed perception-driven, AI-powered QC strategy enhances reproducibility in SSE-3DP.
• Automated inspection and defect identification are enabled through this integrated approach.
• This framework offers a practical pathway toward standardized quality assurance in SSE-3DP.