Publication: Identify target area for spraying using artificial intelligence
| datacite.subject.fos | oecd::Engineering and technology::Electrical engineering, Electronic engineering, Information engineering::Electrical and electronic engineering | |
| dc.contributor.author | Adam Ezrai bin Samurol @ Mohd Faisal | |
| dc.date.accessioned | 2025-05-07T08:04:17Z | |
| dc.date.available | 2025-05-07T08:04:17Z | |
| dc.date.issued | 2023-08 | |
| dc.description.abstract | Implementing convolutional neural network (CNN) in the spray painting industry can increase the efficiency of the spray process and also reduce risk towards people that work for auto body shop that still uses self-spray painting method. This project implements semantic segmentation technique with CNN models to predict and identify the target area that needs to be sprayed. The CNN model was developed by using Google Colab, an Integrated Development Environment (IDE). A dataset was searched online to act as input to train the model. This dataset is a dataset of 211 images of cars searched online and taken in the streets. The images were annotated through makesense.ai to create masks and labels. A Mask R-CNN with ResNet50 backbone model was used to get good prediction on unseen images of a car. The option network training is also needed to be specified to run the training .Thus, the solver used for this training process is the Adam optimizer, the initial learning rate is 0.001 and the number of epoch is 10. The highest Average Precision AP is 0.834 for boundary box detection and 0.421 for segmentation detection. The system for target area for spray painting has been successfully developed. | |
| dc.identifier.uri | https://erepo.usm.my/handle/123456789/21529 | |
| dc.language.iso | en | |
| dc.title | Identify target area for spraying using artificial intelligence | |
| dc.type | Resource Types::text::report::technical report | |
| dspace.entity.type | Publication | |
| oairecerif.author.affiliation | Universiti Sains Malaysia |