Environment sound classification using an attention-based residual neural network

Tripathi, Achyut Mani; Mishra, Aakansha

doi:10.1016/j.neucom.2021.06.031

Cited by 31 publications

(12 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To the best our knowledge, global pooling layers are not investigated in the architectures developed for ESC, and simply average pooling is used in most networks [20], [34], [35]. To gain insights into other pooling strategies, here, we review some of the pooling methods proposed in the visual domain.…”

Section: B Feature Pooling Methodsmentioning

confidence: 99%

“…However, the proposed CNN-D model equipped with SSRP-MS/T achieves higher and comparable accuracies without use of multiple input feature channels, attention mechanisms, or recurrent networks, and too many learnable parameters. PiczakCNN [17] 80.5% 64.9% 4 Log Mel-Delta 31.5 M SoundNet [56] 92.1% 74.2% 8 Raw Data 3.2 M EnvNet-v1 [27] 87.2% 70.8% 7 Raw Data 48.0 M EnvNet-v2 [57] 91.4% 84.9% 13 Raw Data 101.2 M DS-CNN [58] 92.6% 83.1% 9 Raw Data-Log Mel 2.3 M Residual Network [35] 87.3% -19 Log Mel 11.7 M Attention-based Residual Network [35] 92.0% -19 Log Mel 11.9 M Multi-Stream CNN [18] 93.7% 83.5% 16 Raw Data-Log Mel -Attention-based CNN-GRU [54] 94.2% 86.5% 11 Log Mel-Delta -Attention-based CNN-GRU [25] 93.7% 86.1% Figure 6 shows the normalized confusion matrix generated by the proposed CNN-D+SSRP-T model for ESC-50 dataset. It is seen that most classes achieve classification accuracy higher than 80%.…”

Section: B Performance Comparison With State-of-the-artmentioning

confidence: 99%

“…To evaluate the computational complexity, we computed the number of floating point operations (FLOPs) 3 and the inference time of the model when processing a sample audio recording file 4 . Table 2 shows values computed for the proposed model, baseline model of ESC-50 [17], and three other convolutional networks [27], [35], [25] 5 . From Tables 1 and 2, we can see that EnvNet-V2 has improved the accuracy of the baseline model at the cost of considerable increase of FLOPs and inference time (i.e.…”

Section: Analysis Of Computational Complexitymentioning

confidence: 99%

“…Comparing the Grad-CAM visualizations of GAP and SSRP indicates that SSRP not only focuses on the high-energy T-F regions but it also intensifies the responses of low-energy T-F regions. For instance, the focus on low-energy regions at low-frequencies of clock tick sample (mel indices in range of 0-10) or at highfrequencies of crying baby (mel indices in range of [30][31][32][33][34][35][36][37][38][39][40] is increased by SSRP. This denotes that many low-energy T-F regions may contain salient patterns which can…”

Section: E Visual Analysis Of the Proposed Ssrpmentioning

confidence: 99%

See 3 more Smart Citations

Environmental Sound Classification With Low-Complexity Convolutional Neural Network Empowered by Sparse Salient Region Pooling

Seresht

Mohammadi

2023

IEEE Access

View full text Add to dashboard Cite

Environmental Sound Classification (ESC) is an important field in a broad range of applications, such as smart cities, audio surveillance, and health care. Recently, Convolutional Neural Networks (CNNs) have taken the lead from traditional approaches and have produced promising results. However, the achieved improvements are often accompanied by increasing depth, complexity, and size of the network, which prevents their usage in many practical applications. In this work, our goal is to empower a small-size low-complexity CNN model to achieve superior performance. To this end, we concentrate on the importance of global pooling technique, which is less investigated in ESC. In most previous works, models utilize global average pooling layer which does not consider regional saliency, and thus weakens the salient time-frequency regions contributions to the classification, and also to the training of convolutional kernels. We propose a novel global pooling method, called Sparse Salient Region Pooling (SSRP), which computes the channel descriptors using a sparse subset of features, and guides the model to effectively learn from the more salient time-frequency regions. Experimental results demonstrate that the proposed model with only 700K parameters yields accuracies of 86.7% on ESC-50 and 94.8% on ESC-10, which are comparable to that of the state-of-the-art methods. Compared to the baseline model, our model achieves absolute improvement of 21.8% in accuracy on ESC-50, with 98% smaller model size. Our visual analyses show that SSRP intensifies the responses of low-energy regions such that they contribute even more than high-energy regions to the classification of specific sound classes.

show abstract

Section: B Feature Pooling Methodsmentioning

confidence: 99%

Section: B Performance Comparison With State-of-the-artmentioning

confidence: 99%

Section: Analysis Of Computational Complexitymentioning

confidence: 99%

Section: E Visual Analysis Of the Proposed Ssrpmentioning

confidence: 99%

See 2 more Smart Citations

Environmental Sound Classification With Low-Complexity Convolutional Neural Network Empowered by Sparse Salient Region Pooling

Seresht

Mohammadi

2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Fan et al [ 9 ] use the trunk branch of the residual structure to extract features, and mask branches imitate the attention mechanism to add soft weights to the features extracted from the trunk branch to optimize the extracted features and obtain better performance. Tripathi and Mishra [ 10 ] used four two-layer residual blocks to build the network. After the fourth layer, they used attention modules to deal with intra-class inconsistencies, which improved the compactness and increased by 11.50% and 19.50%, respectively, over the benchmark model on the two datasets.…”

Section: Introductionmentioning

confidence: 99%

Underwater Acoustic Target Recognition Based on Attention Residual Network

Bao-xiang

Cui

et al. 2022

Entropy

View full text Add to dashboard Cite

Underwater acoustic target recognition is very complex due to the lack of labeled data sets, the complexity of the marine environment, and the interference of background noise. In order to enhance it, we propose an attention-based residual network recognition method (AResnet). The method can be used to identify ship-radiated noise in different environments. Firstly, a residual network is used to extract the deep abstract features of three-dimensional fusion features, and then a channel attention module is used to enhance different channels. Finally, the features are classified by the joint supervision of cross-entropy and central loss functions. At the same time, for the recognition of ship-radiated noise in other environments, we use the pre-training network AResnet to extract the deep acoustic features and apply the network structure to underwater acoustic target recognition after fine-tuning. The two sets of ship radiation noise datasets are verified, the DeepShip dataset is trained and verified, and the average recognition accuracy is 99%. Then, the trained AResnet structure is fine-tuned and applied to the ShipsEar dataset. The average recognition accuracy is 98%, which is better than the comparison method.

show abstract

Source identification of weak audio signals using attention based convolutional neural network

Presannakumar,

Mohamed

2023

Appl Intell

View full text Add to dashboard Cite

Environment sound classification using an attention-based residual neural network

Cited by 31 publications

References 24 publications

Environmental Sound Classification With Low-Complexity Convolutional Neural Network Empowered by Sparse Salient Region Pooling

Environmental Sound Classification With Low-Complexity Convolutional Neural Network Empowered by Sparse Salient Region Pooling

Underwater Acoustic Target Recognition Based on Attention Residual Network

Source identification of weak audio signals using attention based convolutional neural network

Contact Info

Product

Resources

About