The field of wildlife and environmental monitoring is experiencing significant advancements with the application of machine learning techniques. One of the major challenges in this field is the scarcity of labeled datasets, which is being addressed through innovative data augmentation strategies and active learning approaches. These methods enable the effective training of deep learning models with limited labeled data, leading to improved performance in tasks such as animal detection and semantic segmentation. Furthermore, the generation of synthetic datasets is also gaining traction, particularly in the domain of thermal aerial imaging, where the collection of real-world data is costly and logistically challenging. The development of multi-modal datasets that combine imagery and DNA data is also pushing the boundaries of species classification and instance segmentation. Noteworthy papers in this area include:

• A model-agnostic active learning approach for animal detection from camera traps, which integrates uncertainty and diversity quantities of samples into the active learning sample selection process.
• A novel procedural pipeline for generating synthetic thermal images from an aerial perspective, which enhances existing thermal datasets by introducing new object categories.