MTARSI-INNAR

The MTARSI-INNAR, boasts a meticulous curation process that ensures freedom from cross-contamination. The careful selection of categories (classes) is distinctly highlighted in the accompanying table. This dataset comprises 27 classes allocated to the Training set, each representing a unique aircraft type. Notably, the unseen samples are strategically distributed within the Development set, contributing to a robust training environment. Additionally, the Test set encompasses 16 carefully chosen classes, further underlining the intentional selection process applied to the dataset. This comprehensive approach ensures a balanced representation of diverse aircraft types across training, development, and test sets, facilitating robust analysis and model evaluation. The strategic distribution of samples underscores the dataset's adaptability, making it a valuable resource for researchers and practitioners in the fields of aviation, artificial intelligence, and image analysis.