[Retracted] Identifying Animals in Camera Trap Images via Neural Architecture Search

Figure 8

Examples of misclassified images from ENA24. This figure illustrates four misclassified image samples together with corresponding prediction results. The results associated with each image sample illustrate five top predictions yielded by NASRT, and both the predicted species names and the corresponding probabilities are represented by texts under the image sample. The probabilities are visualized by bars of different colours for visual discrimination. The ground-truth species are highlighted by green texts, and the erroneously predicted species are indicated by red texts. The animals of the ground-truth species in sample images are enclosed by red lines for enhancing the visibility of animals.