A deep learning-based pipeline for mosquito detection and classification from wingbeat sounds

Yin, Myat Su; Haddawy, Peter; Ziemer, Tim; Wetjen, Fabian; Supratak, Akara; Siritanakorn, Worameth; Chantanalertvilai, Tharit; Sriwichai, Patchara; Sa-ngamuang, Chaitawat

doi:10.1007/s11042-022-13367-0

Cited by 12 publications

(4 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Numerous AI, machine learning, and deep learning approaches are also currently being developed to allow the identification, and potentially counting [107], in the field of mosquito species at adult [117][118][119][120][121][122][123][124][125][126][127] and aquatic stages [127], using morphological characteristics, or even wingbeat patterns [126], which could be augmented by DNA barcoding analyses [128]. These approaches might also be combined with the use of UAVs to conduct mosquito vector population surveillance with minimal human resource implications [107].…”

Section: Box 3: Mosquito Sampling and Detection Methodsmentioning

confidence: 99%

Considerations for first field trials of low-threshold gene drive for malaria vector control

Connolly,

Burt,

Christophides

et al. 2024

Malar J

View full text Add to dashboard Cite

Sustainable reductions in African malaria transmission require innovative tools for mosquito control. One proposal involves the use of low-threshold gene drive in Anopheles vector species, where a ‘causal pathway’ would be initiated by (i) the release of a gene drive system in target mosquito vector species, leading to (ii) its transmission to subsequent generations, (iii) its increase in frequency and spread in target mosquito populations, (iv) its simultaneous propagation of a linked genetic trait aimed at reducing vectorial capacity for Plasmodium, and (v) reduced vectorial capacity for parasites in target mosquito populations as the gene drive system reaches fixation in target mosquito populations, causing (vi) decreased malaria incidence and prevalence. Here the scope, objectives, trial design elements, and approaches to monitoring for initial field releases of such gene dive systems are considered, informed by the successful implementation of field trials of biological control agents, as well as other vector control tools, including insecticides, Wolbachia, larvicides, and attractive-toxic sugar bait systems. Specific research questions to be addressed in initial gene drive field trials are identified, and adaptive trial design is explored as a potentially constructive and flexible approach to facilitate testing of the causal pathway. A fundamental question for decision-makers for the first field trials will be whether there should be a selective focus on earlier points of the pathway, such as genetic efficacy via measurement of the increase in frequency and spread of the gene drive system in target populations, or on wider interrogation of the entire pathway including entomological and epidemiological efficacy. How and when epidemiological efficacy will eventually be assessed will be an essential consideration before decisions on any field trial protocols are finalized and implemented, regardless of whether initial field trials focus exclusively on the measurement of genetic efficacy, or on broader aspects of the causal pathway. Statistical and modelling tools are currently under active development and will inform such decisions on initial trial design, locations, and endpoints. Collectively, the considerations here advance the realization of developer ambitions for the first field trials of low-threshold gene drive for malaria vector control within the next 5 years.

show abstract

Section: Box 3: Mosquito Sampling and Detection Methodsmentioning

confidence: 99%

Considerations for first field trials of low-threshold gene drive for malaria vector control

Connolly,

Burt,

Christophides

et al. 2024

Malar J

View full text Add to dashboard Cite

show abstract

“…Therefore, it is unsurprising that they constituted 100% of the parasitic taxa in our review. DL models to identify mosquitoes had accuracy estimates reaching 100% (Kiskin et al., 2020; Yin et al., 2023; Zhang et al., 2017). Strong, user‐friendly, open‐source mobile phone applications have been developed to monitor mosquitoes in real‐time using crowdsourced data.…”

Section: Discussionmentioning

confidence: 99%

From buzzes to bytes: A systematic review of automated bioacoustics models used to detect, classify and monitor insects

Kohlberg,

Myers,

Figueroa

2024

Journal of Applied Ecology

View full text Add to dashboard Cite

Insects play vital ecological roles; many provide essential ecosystem services while others are economically devastating pests and disease vectors. Concerns over insect population declines and expansion have generated a pressing need to effectively monitor insects across broad spatial and temporal scales. A promising approach is bioacoustics, which uses sound to study ecological communities. Despite recent increases in machine learning technologies, the status of emerging automated bioacoustics methods for monitoring insects is not well known, limiting potential applications. To address this gap, we systematically review the effectiveness of automated bioacoustics models over the past four decades, analysing 176 studies that met our inclusion criteria. We describe their strengths and limitations compared to traditional methods and propose productive avenues forward. We found automated bioacoustics models for 302 insect species distributed across nine Orders. Studies used intentional calls (e.g. grasshopper stridulation), by‐products of flight (e.g. bee wingbeats) and indirectly produced sounds (e.g. grain movement) for identification. Pests were the most common study focus, driven largely by weevils and borers moving in dried food and wood. All disease vector studies focused on mosquitoes. A quarter of the studies compared multiple insect families. Our review illustrates that machine learning, and deep learning in particular, are becoming the gold standard for bioacoustics automated modelling approaches. We identified models that could classify hundreds of insect species with over 90% accuracy. Bioacoustics models can be useful for reducing lethal sampling, monitoring phenological patterns within and across days and working in locations or conditions where traditional methods are less effective (e.g. shady, shrubby or remote areas). However, it is important to note that not all insect taxa emit easily detectable sounds, and that sound pollution may impede effective recordings in some environmental contexts. Synthesis and applications: Automated bioacoustics methods can be a useful tool for monitoring insects and addressing pressing ecological and societal questions. Successful applications include assessing insect biodiversity, distribution and behaviour, as well as evaluating the effectiveness of restoration and pest control efforts. We recommend collaborations among ecologists and machine learning experts to increase model use by researchers and practitioners.

show abstract

“…Li et al [66] managed to classify five species of mosquitoes based on their sounds, with a success rate of 73%. Similarly, Yin et al, [67] successfully detected and classified several mosquito species with wingbeat sounds using computational techniques. Folliot et al [68] also monitored pollination by insects and tree use by woodpeckers with acoustics methods and artificial intelligence.…”

Section: (C) Environmental Temperaturementioning

confidence: 99%

Towards acoustic monitoring of bees: wingbeat sounds are related to species and individual traits

Rodríguez Ballesteros,

Desjonquères,

Hevia

et al. 2024

Phil. Trans. R. Soc. B

View full text Add to dashboard Cite

Global pollinator decline urgently requires effective methods to assess their trends, distribution and behaviour. Passive acoustics is a non-invasive and cost-efficient monitoring tool increasingly employed for monitoring animal communities. However, insect sounds remain highly unexplored, hindering the application of this technique for pollinators. To overcome this shortfall and support future developments, we recorded and characterized wingbeat sounds of a variety of Iberian domestic and wild bees and tested their relationship with taxonomic, morphological, behavioural and environmental traits at inter- and intra-specific levels. Using directional microphones and machine learning, we shed light on the acoustic signature of bee wingbeat sounds and their potential to be used for species identification and monitoring. Our results revealed that frequency of wingbeat sounds is negatively related with body size and environmental temperature (between-species analysis), while it is positively related with experimentally induced stress conditions (within-individual analysis). We also found a characteristic acoustic signature in the European honeybee that supported automated classification of this bee from a pool of wild bees, paving the way for passive acoustic monitoring of pollinators. Overall, these findings confirm that insect sounds during flight activity can provide insights on individual and species traits, and hence suggest novel and promising applications for this endangered animal group. This article is part of the theme issue ‘Towards a toolkit for global insect biodiversity monitoring’.

show abstract

A deep learning-based pipeline for mosquito detection and classification from wingbeat sounds

Cited by 12 publications

References 31 publications

Considerations for first field trials of low-threshold gene drive for malaria vector control

Considerations for first field trials of low-threshold gene drive for malaria vector control

From buzzes to bytes: A systematic review of automated bioacoustics models used to detect, classify and monitor insects

Towards acoustic monitoring of bees: wingbeat sounds are related to species and individual traits

Contact Info

Product

Resources

About