Scholarlink

The study on artificially designed bone implants for dental reconstructions, trauma surgery, and cancer therapy has to be refined further, as implant acceptance by the medical community is always a key issue. The present work-study of the fabrication of bone scaffolds and the important parameters required to manufacture bone scaffolds which are properties of composite biomaterials, porosity, pore size, strand diameter, scaffold printing architecture, and compressive strength of bone scaffolds is studied. From research articles, data collection for these parameters and the working of 3D bioplotter has been studied. From observations, 3D bioprinting processes have many problems such as the high cost of bioprinting and bioprinters, lack of full automation, lack of compatible materials, speed of bioprinters, and complex workflow. Therefore, there is a need to bridge the gap between current bioprinting processes and Artificial Intelligence. In this, we considered four different machine learning models which are Hist Gradient Boosting Regressor, Gradient Boosting Regressor, K-neighbors Algorithm, and Random Forest Regressor to analyze the collected data in which we considered, the properties of composite biomaterials, porosity, pore size, strand diameter, scaffold printing architecture as input parameters and compressive strength as an output parameter. From the data, the model gets trained, and by testing the dataset the accuracy of the models is found. From the results, we concluded that the Hist Gradient Boosting Regressor model gave the best-predicted output and we found 95.33% accuracy and mean absolute percentage error (MAPE) is 7.6927 for the model. It means that the model gets trained from the training dataset and can predict future output for required inputs with up to 95.33% accuracy in the current procedure, there is more human interaction in the fabrication process. We were able to decrease human intervention and estimate the compressive strength of bone scaffold before fabricating it with the patient-specific inputs using our approach.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

AI-Assisted Model for Predicting a Compressive Strength of 3D Printed Scaffold Used in Bone Tissue Regeneration

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Bhupesh Sarode,
  • Abhaykumar Kuthe,
  • Bhushan Petkar,
  • Ashutosh Bagde,
  • Subodhkumar Daronde,
  • Gunvanta Dhanuskar

摘要

The study on artificially designed bone implants for dental reconstructions, trauma surgery, and cancer therapy has to be refined further, as implant acceptance by the medical community is always a key issue. The present work-study of the fabrication of bone scaffolds and the important parameters required to manufacture bone scaffolds which are properties of composite biomaterials, porosity, pore size, strand diameter, scaffold printing architecture, and compressive strength of bone scaffolds is studied. From research articles, data collection for these parameters and the working of 3D bioplotter has been studied. From observations, 3D bioprinting processes have many problems such as the high cost of bioprinting and bioprinters, lack of full automation, lack of compatible materials, speed of bioprinters, and complex workflow. Therefore, there is a need to bridge the gap between current bioprinting processes and Artificial Intelligence. In this, we considered four different machine learning models which are Hist Gradient Boosting Regressor, Gradient Boosting Regressor, K-neighbors Algorithm, and Random Forest Regressor to analyze the collected data in which we considered, the properties of composite biomaterials, porosity, pore size, strand diameter, scaffold printing architecture as input parameters and compressive strength as an output parameter. From the data, the model gets trained, and by testing the dataset the accuracy of the models is found. From the results, we concluded that the Hist Gradient Boosting Regressor model gave the best-predicted output and we found 95.33% accuracy and mean absolute percentage error (MAPE) is 7.6927 for the model. It means that the model gets trained from the training dataset and can predict future output for required inputs with up to 95.33% accuracy in the current procedure, there is more human interaction in the fabrication process. We were able to decrease human intervention and estimate the compressive strength of bone scaffold before fabricating it with the patient-specific inputs using our approach.