Automated Synthesis of Hierarchical Deep Learning Cascades for Identifying Visually Similar Objects in UAV Imagery
Technologies, Vol. 14, No. 6, Article 360
Editorial summary
The method builds a sequence of classifiers for visually similar objects in UAV images. It groups classes using feature similarity and observed classification errors. Tests on VisDrone2019 and UAV123 compare the automatically built hierarchy with a single model and manually designed cascades. The best reported F1 score is 94.9%, matching manually designed cascades.
Bibliographic reference
Use the publisher’s record for the citation format required by your journal or organisation.
Dmytro Borovyk, Oleksander Barmak, Pavlo Radiuk, Iurii Krak. "Automated Synthesis of Hierarchical Deep Learning Cascades for Identifying Visually Similar Objects in UAV Imagery". Technologies, Vol. 14, No. 6, Article 360, 2026. https://doi.org/10.3390/technologies14060360
Test a method against your business task.
Describe the decision your system needs to support and the data you have. We assess whether a method fits, define a prototype and agree how to measure its quality, response time and operating cost.