More Like this

1. Human Supervision is Key to Achieving Accurate AI-assisted Wildlife Identifications in Camera Trap Images

   https://doi.org/10.5334/cstp.752  

   Huebner, Sarah E; Palmer, Meredith S; Packer, Craig (January 2024, Citizen Science: Theory and Practice)

   Fortson, Lucy; Crowston, Kevin; Kloetzer, Laure; Ponti, Marisa (Ed.)

   Using public support to extract information from vast datasets has become a popular method for accurately labeling wildlife data in camera trap (CT) images. However, the increasing demand for volunteer effort lengthens the time interval between data collection and our ability to draw ecological inferences or perform data-driven conservation actions. Artificial intelligence (AI) approaches are currently highly effective for species detection (i.e., whether an image contains animals or not) and labeling common species; however, it performs poorly on species rarely captured in images and those that are highly visually similar to one another. To capitalize on the best of human and AI classifying methods, we developed an integrated CT data pipeline in which AI provides an initial pass on labeling images, but is supervised and validated by humans (i.e., a “human-in-the-loop” approach). To assess classification accuracy gains, we compare the precision of species labels produced by AI and HITL protocols to a “gold standard” (GS) dataset annotated by wildlife experts. The accuracy of the AI method was species-dependent and positively correlated with the number of training images. The combined efforts of HITL led to error rates of less than 10% for 73% of the dataset and lowered the error rates for an additional 23%. For two visually similar species, human input resulted in higher error rates than AI. While integrating humans in the loop increases classification times relative to AI alone, the gains in accuracy suggest that this method is highly valuable for high-volume CT surveys. 

   more » « less
2. Soundscape Ecology to Assess Environmental and Anthropogenic Controls on Wildlife Behavior: Environmental Data (Northeastern Alaska and Northwestern Canada, 2017-2023)

   https://doi.org/10.18739/A2ZP3W26F  

   Reinking, Adele K; Liston, Glen E (January 2024, NSF Arctic Data Center)

   This project used cutting-edge soundscape observations and analyses to quantify the influence of changing environmental dynamics and increasing anthropogenic activity on the behavior and phenology of migratory caribou, waterfowl, and songbird communities in Arctic-boreal Alaska and northwestern Canada. We used acoustic and camera-trap monitoring methods to evaluate wildlife responses in novel and non-invasive ways across broad spatial ranges during crucial seasons. Our study combined field observations, modeling, and analyses included (1) soundscape measurements, (2) camera-trap observations, (3) automated soundscape analyses, (4) analyses of camera-trap caribou observations, (5) high-resolution modeling of environmental variables, and (6) statistical analyses of wildlife occupancy, diversity, and phenology. This “Environmental Data” dataset package describes and includes the high-resolution environmental variables used in this study. 

   more » « less
3. 0-MMS: Zero-Shot Multi-Motion Segmentation With A Monocular Event Camera

   https://doi.org/10.1109/ICRA48506.2021.9561755  

   Parameshwara, Chethan M.; Sanket, Nitin J.; Singh, Chahat Deep; Fermuller, Cornelia; Aloimonos, Yiannis (January 2021, 2021 IEEE International Conference on Robotics and Automation (ICRA))

   Segmentation of moving objects in dynamic scenes is a key process in scene understanding for navigation tasks. Classical cameras suffer from motion blur in such scenarios rendering them effete. On the contrary, event cameras, because of their high temporal resolution and lack of motion blur, are tailor-made for this problem. We present an approach for monocular multi-motion segmentation, which combines bottom-up feature tracking and top-down motion compensation into a unified pipeline, which is the first of its kind to our knowledge. Using the events within a time-interval, our method segments the scene into multiple motions by splitting and merging. We further speed up our method by using the concept of motion propagation and cluster keyslices.The approach was successfully evaluated on both challenging real-world and synthetic scenarios from the EV-IMO, EED, and MOD datasets and outperformed the state-of-the-art detection rate by 12%, achieving a new state-of-the-art average detection rate of 81.06%, 94.2% and 82.35% on the aforementioned datasets. To enable further research and systematic evaluation of multi-motion segmentation, we present and open-source a new dataset/benchmark called MOD++, which includes challenging sequences and extensive data stratification in-terms of camera and object motion, velocity magnitudes, direction, and rotational speeds. 

   more » « less
4. BoomBox: An Automated Behavioural Response (ABR) camera trap module for wildlife playback experiments

   https://doi.org/10.1111/2041-210X.13789  

   Palmer, Meredith S.; Wang, Chris; Plucinski, Jacinta; Pringle, Robert M. (January 2022, Methods in Ecology and Evolution)

   Abstract Camera traps (CTs) are a valuable tool in ecological research, amassing large quantities of information on the behaviour of diverse wildlife communities. CTs are predominantly used as passive data loggers to gather observational data for correlational analyses. Integrating CTs into experimental studies, however, can enable rigorous testing of key hypotheses in animal behaviour and conservation biology that are otherwise difficult or impossible to evaluate.We developed the 'BoomBox', an open‐source Arduino‐compatible board that attaches to commercially available CTs to form an Automated Behavioural Response (ABR) system. The modular unit connects directly to the CT’s passive infrared (PIR) motion sensor, playing audio files over external speakers when the sensor is triggered. This creates a remote playback system that captures animal responses to specific cues, combining the benefits of camera trapping (e.g. continuous monitoring in remote locations, lack of human observers, large data volume) with the power of experimental manipulations (e.g. controlled perturbations for strong mechanistic inference).Our system builds on previous ABR designs to provide a cheap (~100USD) and customizable field tool. We provide a practical guide detailing how to build and operate the BoomBox ABR system with suggestions for potential experimental designs that address a variety of questions in wildlife ecology. As proof‐of‐concept, we successfully field tested the BoomBox in two distinct field settings to study species interactions (predator–prey and predator–predator) and wildlife responses to conservation interventions.This new tool allows researchers to conduct a unique suite of manipulative experiments on free‐living species in complex environments, enhancing the ability to identify mechanistic drivers of species' behaviours and interactions in natural systems. 

   more » « less
5. Shed female caribou antlers extend records of calving activity on the Arctic National Wildlife Refuge by millennia

   https://doi.org/10.3389/fevo.2022.1059456  

   Miller, Joshua H.; Wald, Eric J.; Druckenmiller, Patrick (February 2023, Frontiers in Ecology and Evolution)

   Caribou ( Rangifer tarandus ) have among the longest annual migrations of any terrestrial mammal as they move from winter ranges to spring calving grounds. Biomonitoring records indicate broad consistencies in calving geography across the last several decades, but how long have herds used particular calving grounds? Furthermore, how representative are modern patterns of calving geography to periods that pre-date recent climatic perturbations and increased anthropogenic stresses? While modern ecological datasets are not long enough to address these questions, bones from past generations of caribou lying on the tundra provide unique opportunities to study historical calving geography. This is possible because female caribou shed their antlers within days of giving birth, releasing a skeletal indicator of calving. Today, the Coastal Plain of the Arctic National Wildlife Refuge (Alaska) is a key calving ground for the Porcupine Caribou Herd (PCH). To test the duration across which caribou have used this area as a calving ground, we radiocarbon dated three highly weathered female antlers collected from tundra surfaces on the Coastal Plain. Calibrated radiocarbon dates indicate that these antlers were shed between ~1,600 and more than 3,000 calendar years ago. The antiquity of these shed antlers provides the first physical evidence of calving activity on the PCH calving grounds from previous millennia, substantiating the long ecological legacy of the Coastal Plain as a caribou calving ground. Comparisons to published lake core records also reveal that dates of two of the antlers correspond to periods with average summer temperatures that were warmer than has been typical during the last several decades of biomonitoring. This finding expands the range of climatic settings in which caribou are known to use the current PCH calving grounds and suggests that the Coastal Plain of the Arctic Refuge may remain an important caribou calving ground during at least portions of predicted future warming. Discarded skeletal materials provide opportunities to assess the historical states of living populations, including aspects of reproductive biology and migration. Particularly in high-latitude settings, these insights can extend across millennia and offer rare glimpses into the past that can inform current and future management policies. 

   more » « less