Analysing environmental acoustic data through collaboration and automation

Wimmer, Jason; Towsey, Michael; Planitz, Birgit; Williamson, Ian; Roe, Paul

doi:10.1016/j.future.2012.03.004

Cited by 61 publications

(49 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Hence we make use of the following types of test results; A True Positive (TP) (which detects the correct class), A True Negative (TN) (which does not detect the class when it is absent), A False Positive (FP) (which detects the class even if it is absent), and A False Negative (FN) (which detects the wrong class). Using these types of test results we define two performance metric as appear in Table II namely (3) and (4) Recall (Sensitivity) relates to the test's ability to identify a class correctly and Precision (Predictive value positive) is the proportion of positives that correspond to the presence of the class. There may be a problem in identification using this algorithm, when two or more birds sing simultaneously with comparable amount of intensity captured by recording device.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Audio based Bird Species Recognition Using Naïve Bayes Algorithm

2018

IJMTER

View full text Add to dashboard Cite

Abstract-Small of the smallest living creature is the part of the whole ecosystem. Monitoring and maintaining this environmental health has become a task of prime importance, since human activity and climate change has placed the greater pressure on biodiversity. This helps in monitoring dangerous situations to planes caused by birds near airports and also in saving endangered species. Powerful audio signal processing techniques makes it possible to introduce automated methods for the detection of such vocalizations. The complete identification procedure essentially involves collection of recorded vocalizations of different species, audio pre-processing to remove the noise and silence, segmentation to select the most representative characterizing elements of the signal called syllable, extraction of different features commonly found in sound classification and speech recognition to obtain relevant characteristics, and decision procedures for identification, where Machine Learning (ML) algorithms are used to train employed classifiers using labelled database of previously known species of interest. Naïve Bayes algorithm is used for classification and hence recognition of bird species. The required class probabilities are found to be as good as 100% and as worst as 97% in some cases.

show abstract

Section: Resultsmentioning

confidence: 99%

“…Jason Wimmer et al [4] used approach of building comparatively few recognizers capable of recognizing generic features such as oscillations, whistles, whips and stacked harmonics without building a universal classifier. MFCC features followed by HMM were found suitable only for high quality single-syllable calls.…”

Section: Related Workmentioning

confidence: 99%

Audio based Bird Species Recognition Using Naïve Bayes Algorithm

2018

IJMTER

View full text Add to dashboard Cite

show abstract

“…Finally, problems can arise from participants engaging in activities which target popular species, limiting the inferences that can be made about the wider ecosystems [21]. For example, birds and amphibians are considered indicator species for assessing environmental health [6], which creates additional value for their surveying. This brings to question how potential participants' time should be most effectively used when surveying.…”

Section: A Social Implications Of Citizen Sciencementioning

confidence: 99%

“…With this constraint in mind, there is a trend in utilizing the available computational power on smartphones for more localized classification. This type of automated classification is a paradigm shift for many naturalists [6], which highlights the need to better understand the implications and impact of the technology. As a first step towards this, we consider the wider social impact of citizen science technology on both curators and participants.…”

Section: Introductionmentioning

confidence: 99%

Field testing a rare species bioacoustic smartphone application: Challenges and future considerations

Pantidi

Moran

Bachour

et al. 2014

2014 IEEE International Conference on Pervasive Computing and Communication Workshops (PERCOM WORKSHOPS)

View full text Add to dashboard Cite

Abstract-the New Forest cicada is a declining species native to the UK, and the last unconfirmed sighting was in 2000. One of the difficulties in identifying the cicada is that it sings at a high frequency typically inaudible to adults. In this paper we describe a field test of a novel citizen science smartphone application designed to detect and classify the cicada's call. We discuss some of the obstacles to studying this novel technology, and describe the results from a user trial with a simulated cicada. Our observations are then used to inform a series of design considerations for those developing a similar class of application, and improvements for the application itself.

show abstract

“…Researchers can process acoustic data multiple times for varying purposes, whereas information from traditional surveys retains any bias or errors in the original observation (Wimmer, Towsey, Planitz, Williamson, & Roe, 2012). Arrays of acoustic sensors enable direction detection (Blumstein et al, 2011), and the sound recordings are of a reasonable size to store and manage.…”

Section: Introductionmentioning

confidence: 99%

Virtual birding : extending birdwatching to review acoustic recordings

Cottman-Fields¹

View full text Add to dashboard Cite

Observations made by birders have been integral to avian research and conservation efforts for decades. However, broader sources of data about avian activity are needed, to improve the information available for environmental management. Technology advances have seen acoustic monitoring emerge as a new way to study birds. The ever-increasing raw data from acoustic sensors requires processing, to extract information about bird activity. This is a difficult analysis problem, as there are just under 900 bird species in Australia and most have multiple distinct, yet variable, calls.There is an opportunity to extend birding activities to include collection and review of The research methodology was framed around User-centred design and Reflective Agile Iterative Design. The five qualitative research activities used interviews, field studies, exploratory prototype websites, and questionnaires. The results showed that participants invested energy into observing birds at favourite local areas and that their identification skills were region-specific. Many birders were eager to find uncommon birds on bird walks, and in recorded audio from places and times of day that were otherwise difficult to access. Acoustic recordings offered opportunities to enhance birdcall identification skills and find birds that may otherwise be undetected. Birders could apply their acoustic identification knowledge to recorded audio, to identify a range of birdcalls with moderate accuracy, with the aid of spectrogram images: time-frequency-amplitude visualisations of recorded audio. This thesis contributed website interface designs based on birding practices, to enable experienced birders to apply knowledge and skills that are not available elsewhere. These results indicate that integrating scientific acoustic monitoring endeavours into Australian birders' practices can provide expanded and reliable information for avian research and conservation efforts.iv

show abstract

Analysing environmental acoustic data through collaboration and automation

Cited by 61 publications

References 22 publications

Audio based Bird Species Recognition Using Naïve Bayes Algorithm

Audio based Bird Species Recognition Using Naïve Bayes Algorithm

Field testing a rare species bioacoustic smartphone application: Challenges and future considerations

Virtual birding : extending birdwatching to review acoustic recordings

Contact Info

Product

Resources

About