An active learning model to classify animal species in Hong Kong
This work addresses the challenge of applying automated classification models to independently collected camera trap data in specific regions like Hong Kong, which is incremental as it builds on existing active learning methods.
The researchers tackled the problem of automating animal species classification in camera trap images from Hong Kong, using a deep active learning workflow to improve model generalizability for region-specific data.
Camera traps are used by ecologists globally as an efficient and non-invasive method to monitor animals. While it is time-consuming to manually label the collected images, recent advances in deep learning and computer vision has made it possible to automating this process [1]. A major obstacle to this is the generalisability of these models when applying these images to independently collected data from other parts of the world [2]. Here, we use a deep active learning workflow [3], and train a model that is applicable to camera trap images collected in Hong Kong.