Searching for the Ideal Recipe for Preparing Synthetic Data in the Multi-Object Detection Problem
The advancement of deep learning methods across various applications has forced the creation of enormous training datasets. However, obtaining suitable real-world datasets is often challenging for various reasons. Consequently, numerous studies have emerged focusing on the generation and utilization...
Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-01-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/15/1/354 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841549369996738560 |
|---|---|
| author | Michał Staniszewski Aleksander Kempski Michał Marczyk Marek Socha Paweł Foszner Mateusz Cebula Agnieszka Labus Michał Cogiel Dominik Golba |
| author_facet | Michał Staniszewski Aleksander Kempski Michał Marczyk Marek Socha Paweł Foszner Mateusz Cebula Agnieszka Labus Michał Cogiel Dominik Golba |
| author_sort | Michał Staniszewski |
| collection | DOAJ |
| description | The advancement of deep learning methods across various applications has forced the creation of enormous training datasets. However, obtaining suitable real-world datasets is often challenging for various reasons. Consequently, numerous studies have emerged focusing on the generation and utilization of synthetic data in the training process. Hence, there is no universal formula for preparing synthetic data and leveraging it in network training to maximize the effectiveness of various detection methods. This work provides a comprehensive overview of several synthetic data generation techniques, followed by a thorough investigation into the impact of training methods and the selection of synthetic data quantities. The outcomes of this research enable the formulation of conclusions regarding the recipe for developing synthetic data with high efficacy in enhancing detection methods. The main conclusion for the synthetic data generation methods is to ensure maximum diversity at a high level of photorealism, which allows improving the classification quality by more than 5% to even 19% for different detection metrics. |
| format | Article |
| id | doaj-art-af3e615a5b7a4cd3a8f6f4fc12b499d3 |
| institution | Kabale University |
| issn | 2076-3417 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj-art-af3e615a5b7a4cd3a8f6f4fc12b499d32025-01-10T13:15:17ZengMDPI AGApplied Sciences2076-34172025-01-0115135410.3390/app15010354Searching for the Ideal Recipe for Preparing Synthetic Data in the Multi-Object Detection ProblemMichał Staniszewski0Aleksander Kempski1Michał Marczyk2Marek Socha3Paweł Foszner4Mateusz Cebula5Agnieszka Labus6Michał Cogiel7Dominik Golba8Department of Computer Graphics, Vision and Digital Systems, Faculty of Automatic Control, Electronics and Computer Science, Silesian University of Technology, Akademicka 2A, 44-100 Gliwice, PolandDepartment of Computer Graphics, Vision and Digital Systems, Faculty of Automatic Control, Electronics and Computer Science, Silesian University of Technology, Akademicka 2A, 44-100 Gliwice, PolandDepartment of Data Science and Engineering, Faculty of Automatic Control, Electronics and Computer Science, Silesian University of Technology, Akademicka 16, 44-100 Gliwice, PolandDepartment of Data Science and Engineering, Faculty of Automatic Control, Electronics and Computer Science, Silesian University of Technology, Akademicka 16, 44-100 Gliwice, PolandDepartment of Computer Graphics, Vision and Digital Systems, Faculty of Automatic Control, Electronics and Computer Science, Silesian University of Technology, Akademicka 2A, 44-100 Gliwice, PolandDepartment of Computer Graphics, Vision and Digital Systems, Faculty of Automatic Control, Electronics and Computer Science, Silesian University of Technology, Akademicka 2A, 44-100 Gliwice, PolandDepartment of Urban and Spatial Planning, Faculty of Architecture, Silesian University of Technology, Akademicka 7, 44-100 Gliwice, PolandQSystems.pro sp. z o.o. Mochnackiego 34, 41-907 Bytom, PolandQSystems.pro sp. z o.o. Mochnackiego 34, 41-907 Bytom, PolandThe advancement of deep learning methods across various applications has forced the creation of enormous training datasets. However, obtaining suitable real-world datasets is often challenging for various reasons. Consequently, numerous studies have emerged focusing on the generation and utilization of synthetic data in the training process. Hence, there is no universal formula for preparing synthetic data and leveraging it in network training to maximize the effectiveness of various detection methods. This work provides a comprehensive overview of several synthetic data generation techniques, followed by a thorough investigation into the impact of training methods and the selection of synthetic data quantities. The outcomes of this research enable the formulation of conclusions regarding the recipe for developing synthetic data with high efficacy in enhancing detection methods. The main conclusion for the synthetic data generation methods is to ensure maximum diversity at a high level of photorealism, which allows improving the classification quality by more than 5% to even 19% for different detection metrics.https://www.mdpi.com/2076-3417/15/1/354mutli-object detectionsynthetic data generationdeep and transfer learning |
| spellingShingle | Michał Staniszewski Aleksander Kempski Michał Marczyk Marek Socha Paweł Foszner Mateusz Cebula Agnieszka Labus Michał Cogiel Dominik Golba Searching for the Ideal Recipe for Preparing Synthetic Data in the Multi-Object Detection Problem Applied Sciences mutli-object detection synthetic data generation deep and transfer learning |
| title | Searching for the Ideal Recipe for Preparing Synthetic Data in the Multi-Object Detection Problem |
| title_full | Searching for the Ideal Recipe for Preparing Synthetic Data in the Multi-Object Detection Problem |
| title_fullStr | Searching for the Ideal Recipe for Preparing Synthetic Data in the Multi-Object Detection Problem |
| title_full_unstemmed | Searching for the Ideal Recipe for Preparing Synthetic Data in the Multi-Object Detection Problem |
| title_short | Searching for the Ideal Recipe for Preparing Synthetic Data in the Multi-Object Detection Problem |
| title_sort | searching for the ideal recipe for preparing synthetic data in the multi object detection problem |
| topic | mutli-object detection synthetic data generation deep and transfer learning |
| url | https://www.mdpi.com/2076-3417/15/1/354 |
| work_keys_str_mv | AT michałstaniszewski searchingfortheidealrecipeforpreparingsyntheticdatainthemultiobjectdetectionproblem AT aleksanderkempski searchingfortheidealrecipeforpreparingsyntheticdatainthemultiobjectdetectionproblem AT michałmarczyk searchingfortheidealrecipeforpreparingsyntheticdatainthemultiobjectdetectionproblem AT mareksocha searchingfortheidealrecipeforpreparingsyntheticdatainthemultiobjectdetectionproblem AT pawełfoszner searchingfortheidealrecipeforpreparingsyntheticdatainthemultiobjectdetectionproblem AT mateuszcebula searchingfortheidealrecipeforpreparingsyntheticdatainthemultiobjectdetectionproblem AT agnieszkalabus searchingfortheidealrecipeforpreparingsyntheticdatainthemultiobjectdetectionproblem AT michałcogiel searchingfortheidealrecipeforpreparingsyntheticdatainthemultiobjectdetectionproblem AT dominikgolba searchingfortheidealrecipeforpreparingsyntheticdatainthemultiobjectdetectionproblem |