in

Improved pore detection in additive manufacturing using mixup augmentation

Deep learning with mixup augmentation for improved pore detection during additive manufacturing

The study conducted experiments using LPBF at Lawrence Livermore National Laboratory with varying laser powder and speed settings. Acoustic measurements were recorded during the process using a microphone placed in the build chamber. Data was collected at a sampling rate of 100 kHz and analyzed to identify pores in the material. X-ray radiography was used to locate pores ex-situ, and the acoustic data was aligned with these pore locations for analysis. The data was divided into pore and non-pore segments for classification.

To address class imbalance and enhance model performance, Mixup data augmentation was employed. This technique creates augmented data by combining samples with different weights, improving generalization and reducing overfitting. The augmented data was used to train a Convolutional Neural Network (CNN) to classify acoustic signals as pores or non-pores. The CNN architecture included convolutional layers, max-pooling, and a dense layer for classification.

The performance of the CNN model was evaluated using metrics such as accuracy, precision, recall, and F1 score. Precision measures the proportion of correctly identified pores, recall measures the proportion of actual pores correctly identified, and the F1 score is a combination of precision and recall. These metrics were used to assess the model’s ability to classify pores in the acoustic data accurately.

Source link

Source link: https://www.nature.com/articles/s41598-024-63288-1

What do you think?

Leave a Reply

GIPHY App Key not set. Please check settings

The SHOCKING Deal: Apple Joins Forces with OpenAI and Musk is Not Happy!

#Apple, OpenAI, and Musk: A Surprising Collaboration Unveiled #Innovation

LinkedIn’s AI job coach can write your cover letters and edit your resumé

LinkedIn’s AI job coach can write cover letters, edit resumé. #AIcareercoach