Plant and Plant Disease Identification
We have been working on plant identification from photographs since 2008 and also participating in the CLEF plant recognition campaigns since 2011 with 1st and 2nd place results in 2012 and 2016. The work is joint with my students and Dr. Erchan Aptoula from Gebze Technical University.
Our deep learning ensemble is able to identify the plant in an unconstrained photograph with around 75% accuracy among 10,000 plant species. This accuracy matches the level of some of the human experts at the campaign! You can test our plant identification system at plantid.sabanciuniv.edu
More recently we have also started working on plant disease identification with the goal of detecting diseases and deficiencies early on.
á Sara Atito, Berrin A. Yanikoglu, Erchan Aptoula, Ipek Ganiyusufoglu, Aras Yildiz, Kerem Yildirir, Baris Sevilmis, M. Umut Sen: Plant Identification with Deep Learning Ensembles. CLEF (Working Notes) 2018
á Mostafa Mehdipour-Ghazi, Berrin A. Yanikoglu, Erchan Aptoula: Plant identification using deep neural networks via optimization of transfer learning parameters. Neurocomputing 235: 228-235 (2017) pdf
Results of the 2018 campaign with 10,000 classes:
Results of the 2016 campaign with 1,000 classes: