Real-Time Product Label Classification in a Manufacturing Environment
(2024) DABN01 20241Department of Statistics
Department of Economics
- Abstract
- In this thesis, a convolutional neural network was designed to perform real-time image classifications of ten product classes within a fast-paced, unstable manufacturing environment for quality control, while also linking it to a business case. Out of six eligible models that performed well on test data, only one proved good performance in simulated test runs, using out-of-sample test data. The find ings of this thesis highlights the necessity of maintaining a critical perspective on results obtained from test data during model evaluations within a manufacturing environment and that additional tests for validating initial results might be required. Furthermore, this thesis presents a framework for implementing computer vision in a... (More)
- In this thesis, a convolutional neural network was designed to perform real-time image classifications of ten product classes within a fast-paced, unstable manufacturing environment for quality control, while also linking it to a business case. Out of six eligible models that performed well on test data, only one proved good performance in simulated test runs, using out-of-sample test data. The find ings of this thesis highlights the necessity of maintaining a critical perspective on results obtained from test data during model evaluations within a manufacturing environment and that additional tests for validating initial results might be required. Furthermore, this thesis presents a framework for implementing computer vision in a manufacturing environment. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9163547
- author
- Nederlund Persson, Oliver LU and Olsson, Anton LU
- supervisor
- organization
- course
- DABN01 20241
- year
- 2024
- type
- H1 - Master's Degree (One Year)
- subject
- keywords
- Computer vision, manufacturing, convolutional neural networks
- language
- English
- id
- 9163547
- date added to LUP
- 2024-09-25 14:43:16
- date last changed
- 2024-09-25 14:43:16
@misc{9163547, abstract = {{In this thesis, a convolutional neural network was designed to perform real-time image classifications of ten product classes within a fast-paced, unstable manufacturing environment for quality control, while also linking it to a business case. Out of six eligible models that performed well on test data, only one proved good performance in simulated test runs, using out-of-sample test data. The find ings of this thesis highlights the necessity of maintaining a critical perspective on results obtained from test data during model evaluations within a manufacturing environment and that additional tests for validating initial results might be required. Furthermore, this thesis presents a framework for implementing computer vision in a manufacturing environment.}}, author = {{Nederlund Persson, Oliver and Olsson, Anton}}, language = {{eng}}, note = {{Student Paper}}, title = {{Real-Time Product Label Classification in a Manufacturing Environment}}, year = {{2024}}, }