AnimalFinder: A semi-automated system for animal detection in time-lapse camera trap images

Tack, Jennifer L. Price; West, Brian S.; McGowan, Conor P.; Ditchkoff, Stephen S.; Reeves, Stanley J.; Keever, Allison C.; Grand, James B.

doi:10.1016/j.ecoinf.2016.11.003

Cited by 42 publications

(23 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Even motion triggered camera traps suffer from many false‐positive images due to wind and moving vegetation. In computer vision, finding novel objects within series of images can be achieved using background subtraction, which distinguishes sedentary objects, such as trees and clouds, from moving objects, such as animals, within videos or groups of images (Price Tack et al., ; Ren, Han, & He, ; Weinstein, ) (Figure a). A background model is created by computing an expected image based on the previous pixel values (Stauffer & Grimson, ).…”

Section: Countingmentioning

confidence: 99%

“…These studies report high accuracy in removing empty frames, but there were persistent challenges in reducing false positives from strong wind and other extraneous movement in heterogeneous environments (Price Tack et al. ). Tailoring detection algorithms to individual taxa can greatly improve accuracy, for example, Zeppelzauer () reported >95% accuracy in detecting African elephants ( Loxodonta cyclotis ) by building a colour model from training data.…”

Section: Countingmentioning

confidence: 99%

“…The background model changes over time based on new pixel values, thereby reducing false positives from shifts in illumination and external movement, such as wind, waves or camera shake. Once images have been divided into foreground and background pixels (known as segmentation), objects are partitioned into discrete groups, with connected sets of pixels corresponding to individual organisms.I found 55 articles that used a form of background subtraction to detect and count animals, primarily for mammals (n = 24) and birds (n = 22).These studies report high accuracy in removing empty frames, but there were persistent challenges in reducing false positives from strong wind and other extraneous movement in heterogeneous environments (PriceTack et al 2016). Tailoring detection algorithms to individual taxa can greatly improve accuracy, for example, Zeppelzauer (2013) reportedF I G U R E 3 Applications of computer vision to describing ecological objects.…”

mentioning

confidence: 99%

See 2 more Smart Citations

A computer vision for animal ecology

Weinstein

2017

Journal of Animal Ecology

342

287

View full text Add to dashboard Cite

A central goal of animal ecology is to observe species in the natural world. The cost and challenge of data collection often limit the breadth and scope of ecological study. Ecologists often use image capture to bolster data collection in time and space. However, the ability to process these images remains a bottleneck. Computer vision can greatly increase the efficiency, repeatability and accuracy of image review. Computer vision uses image features, such as colour, shape and texture to infer image content. I provide a brief primer on ecological computer vision to outline its goals, tools and applications to animal ecology. I reviewed 187 existing applications of computer vision and divided articles into ecological description, counting and identity tasks. I discuss recommendations for enhancing the collaboration between ecologists and computer scientists and highlight areas for future growth of automated image analysis.

show abstract

Section: Countingmentioning

confidence: 99%

Section: Countingmentioning

confidence: 99%

mentioning

confidence: 99%

See 1 more Smart Citation

A computer vision for animal ecology

Weinstein

2017

Journal of Animal Ecology

342

287

View full text Add to dashboard Cite

show abstract

“…By defining the image features for a background model, background subtraction can be used to filter out empty frames. The increase in accessibility of computer vision libraries (Bradski, 2000) has led to a hope for a single background subtraction method for diverse ecological systems (Bowley, Andes, Ellis-Felege, & Desell, 2016;Elias, Golubovic, Krintz, & Wolski, 2017;Price Tack et al, 2016;Weinstein, 2015). However, the complex backgrounds of marine and terrestrial environments, combined with the myriad forms of animal shape and patterning, make a universal detector unlikely.…”

mentioning

confidence: 99%

“…My goal is to explore a pipeline for videobased biodiversity observation, describe its strengths and limitations, and outline avenues for further exploration. Previous work in ecological computer vision has used motion detection to screening images based on temporal filtering (Swinnen, Reijniers, Breno, & Leirs, 2014), pixel models (Weinstein, 2015), and edge detection (Price Tack et al, 2016). I build on background subtraction tools introduced in Weinstein (2015) to add deep learning classification of potential movement objects.…”

mentioning

confidence: 99%

Scene‐specific convolutional neural networks for video‐based biodiversity detection

Weinstein

2018

Methods Ecol Evol

View full text Add to dashboard Cite

Finding, counting and identifying animals is a central challenge in ecology. Most studies are limited by the time and cost of fieldwork by human observers. To increase the spatial and temporal breadth of sampling, ecologists are adopting passive image‐based monitoring approaches. While passive monitoring can expand data collection, a remaining obstacle is finding the small proportion of images containing ecological objects among the majority of frames containing only background scenes. I proposed a scene‐specific convolutional neural network for detecting animals of interest within long duration time‐lapse videos. Convolutional neural networks are a type of deep learning algorithm that have recently made significant advances in image classification. The approach was tested on videos of floral visitation by hummingbirds. Despite low frame rates, poor image quality, and complex video conditions, the model correctly classified over 90% of frames containing hummingbirds. Combining motion detection and image classification can substantially reduce the time investment in scoring images from passive monitoring studies. These results underscore the promise of deep learning to lead ecology into greater automation using passive image analysis. To help facilitate future applications, I created a desktop executable that can be used to apply pre‐trained models to videos, as well as reproducible scripts for training new models on local and cloud environments.

show abstract

Potentiality and limitations of N‐mixture and Royle‐Nichols models to estimate animal abundance based on noninstantaneous point surveys

Nakashima

2019

Population Ecology

View full text Add to dashboard Cite

Reliable and accurate information on animal abundance is fundamental for the conservation and management of wildlife. Recently, a number of innovative devices (such as camera traps) have been widely used in field surveys and have largely improved survey efficiency. However, these devices often constitute noninstantaneous point surveys, resulting in the multiple counts of the same animal individuals within a single sampling occasion (i.e., false-positive errors). Many commonly-used statistical models do not explicitly account for the false-positive error, with its effects on estimates being poorly understood. Here, I tested the performance of the commonly-used Poisson-binomial N-mixture and the Royle-Nichols model in the presence of both false-positive and negative errors (i.e., individuals in a population might not be detected). I also implemented the Poisson-Poisson mixture model in the Bayesian framework to evaluate its reliability. The results of the simulation using random walks based on Ornstein-Uhlenbeck processes showed that the Poisson-binomial model was not robust to false-positive errors. In comparison, the Royle-Nichols and Poisson-Poisson models provided reasonable estimates of the number of animals whose home range included the survey point. However, the number of animals whose home range included the survey point is inherently influenced by the size of animal home ranges, and thus cannot be used as a surrogate of animal density. Although the N-mixture and Royle-Nichols models are widely used, their utility might be restricted by this limitation. In conclusion, studies should clearly define the objective of surveys and carefully consider whether the models used are valid. K E Y W O R D Scamera trap, detection probability, home range, N-mixture model, Royle-Nichols model

show abstract

AnimalFinder: A semi-automated system for animal detection in time-lapse camera trap images

Cited by 42 publications

References 25 publications

A computer vision for animal ecology

A computer vision for animal ecology

Scene‐specific convolutional neural networks for video‐based biodiversity detection

Potentiality and limitations of N‐mixture and Royle‐Nichols models to estimate animal abundance based on noninstantaneous point surveys

Contact Info

Product

Resources

About