Exportar Publicação

Marson, L., Costa, H., Leithardt, V. & Crocker, P. (2025). Detecting faults in furniture parts using neural networks in a mobile app. In Dumitrache I., Moisescu M., Pop F., Copot D., Popescu D. (Ed.), 2025 25th International Conference on Control Systems and Computer Science (CSCS), Proceedings (pp. 306?313). IEEE. https://doi.org/10.1109/CSCS66924.2025.00052

L. M. Marson et al.,  "Detecting faults in furniture parts using neural networks in a mobile app", in 2025 25th Int. Conf. on Control Systems and Computer Science (CSCS), Proc. , Dumitrache I., Moisescu M., Pop F., Copot D., Popescu D., Ed., Bucharest, Romania, IEEE, 2025, pp. 306-313

@inproceedings{marson2025_1791476466020,
	author = "Marson, L. and Costa, H. and Leithardt, V. and Crocker, P.",
	title = "Detecting faults in furniture parts using neural networks in a mobile app",
	booktitle = "2025 25th International Conference on Control Systems and Computer Science (CSCS), Proceedings ",
	year = "2025",
	editor = "Dumitrache I., Moisescu M., Pop F., Copot D., Popescu D.",
	volume = "",
	number = "",
	series = "",
	doi = "10.1109/CSCS66924.2025.00052",
	pages = "306-313",
	publisher = "IEEE",
	address = "Bucharest, Romania",
	organization = "",
	url = "https://ieeexplore.ieee.org/document/11181590"
}

TY  - CPAPER
TI  - Detecting faults in furniture parts using neural networks in a mobile app
T2  - 2025 25th International Conference on Control Systems and Computer Science (CSCS), Proceedings 
AU  - Marson, L.
AU  - Costa, H.
AU  - Leithardt, V.
AU  - Crocker, P.
PY  - 2025
SP  - 306-313
SN  - 2379-0474
DO  - 10.1109/CSCS66924.2025.00052
CY  - Bucharest, Romania
UR  - https://ieeexplore.ieee.org/document/11181590
AB  - The furniture industry faces significant challenges in detecting defects in parts after the painting process, which are often difficult to identify visually but compromise the final quality of the product. This project aims to develop an application for Android devices that use Deep Learning techniques to detect anomalies in real-time from images of painted parts. The initial process included a comprehensive review of convolutional neural networks and Android application development, as well as discussions with industry professionals to understand the types of common defects. The methodology involved segmenting compliant and non-compliant parts, preprocessing the images, and training a neural network model in Python using TensorFlow and Keras. The trained model was adapted for mobile devices with TensorFlow Lite and integrated into an Android application using ML Kit. The project not only addresses technical fault detection but also explores the viability and applicability of the solution to improve the quality and efficiency of inspection processes in the furniture industry.
ER  -