Bird vocalizations are pivotal for ecological monitoring,providing insights into biodiversity and ecosystem health.Traditional recognition methods often neglect phase information,resulting in incomplete feature repres...Bird vocalizations are pivotal for ecological monitoring,providing insights into biodiversity and ecosystem health.Traditional recognition methods often neglect phase information,resulting in incomplete feature representation.In this paper,we introduce a novel approach to bird vocalization recognition(BVR)that integrates both amplitude and phase information,leading to enhanced species identification.We propose MHARes Net,a deep learning(DL)model that employs residual blocks and a multi-head attention mechanism to capture salient features from logarithmic power(POW),Instantaneous Frequency(IF),and Group Delay(GD)extracted from bird vocalizations.Experiments on three bird vocalization datasets demonstrate our method's superior performance,achieving accuracy rates of 94%,98.9%,and 87.1%respectively.These results indicate that our approach provides a more effective representation of bird vocalizations,outperforming existing methods.This integration of phase information in BVR is innovative and significantly advances the field of automatic bird monitoring technology,offering valuable tools for ecological research and conservation efforts.展开更多
Birds play a crucial role in maintaining ecological balance,making bird recognition technology a hot research topic.Traditional recognition methods have not achieved high accuracy in bird identification.This paper pro...Birds play a crucial role in maintaining ecological balance,making bird recognition technology a hot research topic.Traditional recognition methods have not achieved high accuracy in bird identification.This paper proposes an improved ResNet18 model to enhance the recognition rate of local bird species in Yunnan.First,a dataset containing five species of local birds in Yunnan was established:C.amherstiae,T.caboti,Syrmaticus humiae,Polyplectron bicalcaratum,and Pucrasia macrolopha.The improved ResNet18 model was then used to identify these species.This method replaces traditional convolution with depth wise separable convolution and introduces an SE(Squeeze and Excitation)module to improve the model’s efficiency and accuracy.Compared to the traditional ResNet18 model,this improved model excels in implementing a wild bird classification solution,significantly reducing computational overhead and accelerating model training using low-power,lightweight hardware.Experimental analysis shows that the improved ResNet18 model achieved an accuracy of 98.57%,compared to 98.26%for the traditional Residual Network 18 layers(ResNet18)model.展开更多
基金supported by the Beijing Natural Science Foundation (5252014)the National Natural Science Foundation of China (62303063)。
文摘Bird vocalizations are pivotal for ecological monitoring,providing insights into biodiversity and ecosystem health.Traditional recognition methods often neglect phase information,resulting in incomplete feature representation.In this paper,we introduce a novel approach to bird vocalization recognition(BVR)that integrates both amplitude and phase information,leading to enhanced species identification.We propose MHARes Net,a deep learning(DL)model that employs residual blocks and a multi-head attention mechanism to capture salient features from logarithmic power(POW),Instantaneous Frequency(IF),and Group Delay(GD)extracted from bird vocalizations.Experiments on three bird vocalization datasets demonstrate our method's superior performance,achieving accuracy rates of 94%,98.9%,and 87.1%respectively.These results indicate that our approach provides a more effective representation of bird vocalizations,outperforming existing methods.This integration of phase information in BVR is innovative and significantly advances the field of automatic bird monitoring technology,offering valuable tools for ecological research and conservation efforts.
文摘Birds play a crucial role in maintaining ecological balance,making bird recognition technology a hot research topic.Traditional recognition methods have not achieved high accuracy in bird identification.This paper proposes an improved ResNet18 model to enhance the recognition rate of local bird species in Yunnan.First,a dataset containing five species of local birds in Yunnan was established:C.amherstiae,T.caboti,Syrmaticus humiae,Polyplectron bicalcaratum,and Pucrasia macrolopha.The improved ResNet18 model was then used to identify these species.This method replaces traditional convolution with depth wise separable convolution and introduces an SE(Squeeze and Excitation)module to improve the model’s efficiency and accuracy.Compared to the traditional ResNet18 model,this improved model excels in implementing a wild bird classification solution,significantly reducing computational overhead and accelerating model training using low-power,lightweight hardware.Experimental analysis shows that the improved ResNet18 model achieved an accuracy of 98.57%,compared to 98.26%for the traditional Residual Network 18 layers(ResNet18)model.
