Abstract

The limited availability of information on Chilean nativeflora has resulted in a lackof knowledge among the general public, and the classification of these plantsposes challenges without extensive expertise. This study evaluates theperformance of several Deep Learning(DL) models, namely InceptionV3,VGG19, ResNet152, and MobileNetV2, in classifying images representingChilean nativeflora. The models are pre-trained on Imagenet. A datasetcontaining 500 images for each of the 10 classes of nativeflowers in Chile wascurated, resulting in a total of 5000 images. The DL models were applied to thisdataset, and their performance was compared based on accuracy and otherrelevant metrics. Thefindings highlight the potential of DL models to accuratelyclassify images of Chilean nativeflora. The results contribute to enhancing theunderstanding of these plant species and fostering awareness among the generalpublic. Further improvements and applications of DL in ecology and biodiversityresearch are discussed.