This paper presents a novel approach for accurate counting and localization of tropical plants in aerial images that is able to work in new visual domains in which the available data is not labeled. Our approach uses deep learning and domain adaptation, designed to handle domain shift between the training and test data, which is a common challenge in this agricultural applications. This method uses a source dataset with annotated plants and a target dataset without annotations, and adapts a model trained on the source dataset to the target dataset using unsupervised domain alignment and pseudolabeling. The experimental results show the effectiveness of this approach for plant counting in aerial images of pineapples under significative domain shift, achieving a reduction up to 97% in the counting error when compared to the supervised baseline.

Human Rhinovirus (HRV) is the most common cause of upper respiratory infections and exacerbations of asthma. In this work, we selected 14 peptides (6 from HRV A and 8 from HRV C) encompassing potential CD4 T cell epitopes. Peptides were selected for being highly conserved in HRV A and C serotypes and predicted to bind to multiple HLA II molecules. We found positive T cell recall responses by IFNγ-ELISPOT assays to 8 peptides, validating 7 of them (3 from HRV A and 4 from HRV C) as CD4 T cell epitopes through intracellular cytokine staining assays. Additionally, we verified their promiscuous binding to multiple HLA II molecules by quantitative binding assays. According to their experimental HLA II binding profile, the combination of all these 7 epitopes could be presented and recognized by > 95 % of the world population. We actually determined IFNγ responses to a pool encompassing these CD4 T cell epitopes by intracellular cytokine staining, finding positive responses in 29 out of 30 donors. The CD4 T cell epitopes identified in this study could be key to monitor HRV infections and to develop peptide-based vaccines against most HRV A and C serotypes.