English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

Automated species identification and delimitation is challenging,particularly in rare and thus often scarcely sampled species, which do notallow sufficient discrimination of infraspecific versus interspecificvariation. Typical problems arising from either low or exaggeratedinterspecific morphological differentiation are best met by automated methodsof machine learning that learn efficient and effective species identificationfrom training samples. However, limited infraspecific sampling remains a keychallenge also in machine learning. 1In this study, we assessed whether atwo-level data augmentation approach may help to overcome the problem of scarcetraining data in automated visual species identification. The first level ofvisual data augmentation applies classic approaches of data augmentation andgeneration of faked images using a GAN approach. Descriptive feature vectorsare derived from bottleneck features of a VGG-16 convolutional neural network(CNN) that are then stepwise reduced in dimensionality using Global AveragePooling and PCA to prevent overfitting. The second level of data augmentationemploys synthetic additional sampling in feature space by an oversamplingalgorithm in vector space (SMOTE). Applied on two challenging datasets ofscarab beetles (Coleoptera), our augmentation approach outperformed anon-augmented deep learning baseline approach as well as a traditional 2Dmorphometric approach (Procrustes analysis).

Image-based Automated Species Identification: Can Virtual Data Augmentation Overcome Problems of Insufficient Sampling?