Automated String Art Creation: Integrated Advanced Computational Techniques and Precision Art Designing
The Thread Art Machine project automates the traditional, labour-intensive process of string art creation by integrating advanced computational and manufacturing techniques. Utilizing CAD models in Fusion 360, CNC machining, 3D printing, and Arduino programming, the machine precisely sets threads an...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10844087/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The Thread Art Machine project automates the traditional, labour-intensive process of string art creation by integrating advanced computational and manufacturing techniques. Utilizing CAD models in Fusion 360, CNC machining, 3D printing, and Arduino programming, the machine precisely sets threads and nails to form intricate string art patterns. Key components of the machine include adaptive algorithms for real-time thread tension adjustments and transposed convolution layers for image processing, which enhance both the accuracy and aesthetic quality of the final artwork. The project faced challenges such as material selection, CAD design, and hardware-software interfacing, all of which were addressed through iterative design and validation processes. A convolutional neural network (CNN) was employed to process grayscale images, extracting and reconstructing features using pooling and deconvolution techniques, with the model achieving stable performance over multiple epochs. The machine’ s calibration system, involving LDR sensors and laser alignment, ensures precision in thread placement. The resulting string art pieces, derived from famous images, demonstrate the machine’ s capability to capture key features through varying string densities, offering a blend of mathematical precision and artistic abstraction. Validation through training and testing revealed consistent performance, with minimal overfitting, as indicated by flat training and validation loss and accuracy curves. |
|---|---|
| ISSN: | 2169-3536 |