Bird Image Classification using Convolutional Neural Network Transfer Learning Architectures

International Journal of Advanced Computer Science and Applications, Год журнала: 2023, Номер 14(3)

Опубликована: Янв. 1, 2023

With the technological progress of human beings, more and animal bird species are being endangered sometimes even going to verge extinction. However, existence birds is highly beneficial for civilization as help in pollination, destroying harmful insects crops, etc. To ensure healthy co-existence all along with almost advanced countries have taken up some conservation measures species. conservation, first step identify found different locations. Deep learning-based techniques best suited automated identification from captured images. In this paper, a Convolutional Neural Network based image methodology has been proposed. Four transfer architectures, namely Resnet152V2, Inception V3, Densenet201, MobileNetV2 used classification identification. The models trained using 58388 images belonging 400 birds, tested 2000 birds. Out these four models, Resnet152V2 DenseNet201 performed comparatively well. accuracy was highest at 95.45%, but it faced large loss 0.8835. But on results, though had an 95.05%, less i.e., 0.6854. results show that model can further be real-life classification.

Язык: Английский

---

A Critical Review of Bioaccumulation and Biotransformation of Organic Chemicals in Birds

Reviews of Environmental Contamination and Toxicology, Год журнала: 2022, Номер 260(1)

Опубликована: Май 20, 2022

Abstract A literature review of bioaccumulation and biotransformation organic chemicals in birds was undertaken, aiming to support scoping prioritization future research. The objectives were characterize available bioaccumulation/biotransformation data, identify knowledge gaps, determine how extant data can be used, explore the strategy steps forward. An intermediate approach balanced between expediency rigor taken given vastness literature. Following a critical > 500 peer-reviewed studies, 25,000 entries 2 million information bytes compiled on 700 compounds for ~ 320 wild species 60 domestic breeds birds. These organized into themed databases , field survey microsomal enzyme activity metabolic pathway bird taxonomy diet . Significant gaps identified all at multiple levels. Biotransformation characterization largely fragmented over metabolite/pathway identification or kinetics. Limited kinetic constrained development an avian model. substantial shortage vivo kinetics has been observed as most reported rate constants derived vitro. No metric comprehensively captured key contaminant classes chemical groups broad-scope modeling biotransformation. However, metrics such biota-feed accumulation factor, maximum transfer total elimination constant more readily usable benchmarking than other reviewed parameters. Analysis demonstrated lack shorebirds, seabirds, raptors. In study birds, this revealed need greater diversity, measurements environmental media, basic biometrics exposure conditions, tissues/matrices sampling, further exploration Limitations classical current research strategies used studies also discussed. Forward-looking proposed: adopting roadmap investigations, integrating existing biomonitoring gap-filling with non-testing approaches, improving reporting practices, expanding sampling scopes, bridging models theories, exploring via genomics, establishing online repository.

Язык: Английский

---

Common seed dispersers contribute most to the persistence of a fleshy-fruited tree

Communications Biology, Год журнала: 2023, Номер 6(1)

Опубликована: Март 27, 2023

Abstract Mutualistic interactions are by definition beneficial for each contributing partner. However, it is insufficiently understood how mutualistic influence partners throughout their lives. Here, we used animal species-explicit, microhabitat-structured integral projection models to quantify the effect of seed dispersal 20 species on full life cycle tree Frangula alnus in Białowieża Forest, Eastern Poland. Our analysis showed that increased population growth 2.5%. The effectiveness animals as dispersers was strongly related interaction frequency but not quality dispersal. Consequently, projected decline due simulated extinction driven loss common rather than rare mutualist species. results support notion frequently interacting mutualists contribute most persistence populations partners, underscoring role ecosystem functioning and nature conservation.

Язык: Английский

---

Chronic exposure to tebuconazole impairs offspring growth and survival in farmland birds: an experiment in captive house sparrows.

Environmental Research, Год журнала: 2025, Номер unknown, С. 121321 - 121321

Опубликована: Март 1, 2025

Язык: Английский

---

How Do Bird Population Trends Relate to Human Pressures Compared to Economic Growth?

Sustainability, Год журнала: 2025, Номер 17(8), С. 3506 - 3506

Опубликована: Апрель 14, 2025

Biodiversity loss is a global environmental concern, mainly driven by human-induced factors, encompassing both direct and indirect drivers. This study investigates the long-term relationship between either Human Footprint Index (HFI), which measures extent of human pressures (i.e., drivers), or Gross Domestic Product (GDP), measure economic growth driver) biodiversity change, using bird population trends as indicators. The analysis was based on time-series data for Portugal (2004–2023) aggregated at national sub-national scales, representative different socio-economic contexts. Multi-species indices were regressed against HFI GDP Autoregressive Distributed Lag (ARDL) to identify long-run relationships. Bird varied species group (common, agricultural, forest birds) context underscoring importance assessments. had varying predictive value across groups contexts, with showing greater consistency, particularly predictor agricultural birds. While most models showed negative association abundance GDP, revealing signal populations some suggested mixed results, indicating that conservation policies must take local contexts into account.

Язык: Английский

---