A Paper Base on Synthetic Dataset and Deep Learning Got Published on MDPI Machine Learning and Knowledge Extraction Journal

The paper titled “Deep Learning Based Objection Recognition Using Physically-Realistic Synthetic Depth Scenes” (Baimukashev, D.; Zhilisbayev, A.; Kuzdeuov, A.; Oleinikov, A.; Fadeyev, D.; Makhataeva, Z.; Varol, H.A.) was recently accepted for publication on Machine Learning and Knowledge Extraction, MDPI journal.

In this study, a deep object recognition framework using synthetic depth image dataset was developed and validated. The developed framework can be trained on synthetically generated depth data and then be employed on a real depth dataset in a cluttered environment for object detection. This has the potential to substantially decrease the time and human effort required for the extensive data collection and labeling.