A Convolutional Neural Network model based on Neutrosophy for Noisy Speech Recognition

Convolutional neural networks are sensitive to unknown noisy condition in the test phase and so their performance degrades for the noisy data classification task including noisy speech recognition. In this research, a new convolutional neural network (CNN) model with data uncertainty handling; referred as NCNN (Neutrosophic Convolutional Neural Network); is proposed for classification task. Here, speech signals are used as input data and their noise is modeled as uncertainty. In this task, using speech spectrogram, a definition of uncertainty is proposed in neutrosophic (NS) domain. Uncertainty is computed for each Time-frequency point of speech spectrogram as like a pixel. Therefore, uncertainty matrix with the same size of spectrogram is created in NS domain. In the next step, a two parallel paths CNN classification model is proposed. Speech spectrogram is used as input of the first path and uncertainty matrix for the second path. The outputs of two paths are combined to compute the final output of the classifier. To show the effectiveness of the proposed method, it has been compared with conventional CNN on the isolated words of Aurora2 dataset. The proposed method achieves the average accuracy of 85.96 in noisy train data. It is more robust against noises with accuracies 90, 88 and 81 in test sets A, B and C, respectively. Results show that the proposed method outperforms conventional CNN with the improvement of 6, 5 and 2 percentage in test set A, test set B and test sets C, respectively. It means that the proposed method is more robust against noisy data and handle these data effectively.

[1]  Yanhui Guo,et al.  An Efficient Image Segmentation Algorithm Using Neutrosophic Graph Cut , 2017, Symmetry.

[2]  Graham Currie,et al.  Prediction intervals to account for uncertainties in neural network predictions: Methodology and application in bus travel time prediction , 2011, Eng. Appl. Artif. Intell..

[3]  Abdolreza Rashno,et al.  Content-based image retrieval with color and texture features in neutrosophic domain , 2017, 2017 3rd International Conference on Pattern Recognition and Image Analysis (IPRIA).

[4]  Yanhui Guo,et al.  Color texture image segmentation based on neutrosophic set and wavelet transformation , 2011, Comput. Vis. Image Underst..

[5]  S. Roberts,et al.  Confidence Intervals and Prediction Intervals for Feed-Forward Neural Networks , 2001 .

[6]  M. Gity,et al.  Segmentation of breast ultrasound images based on active contours using neutrosophic theory , 2018, Journal of Medical Ultrasonics.

[7]  Yanmin Qian,et al.  Very Deep Convolutional Neural Networks for Noise Robust Speech Recognition , 2016, IEEE/ACM Transactions on Audio, Speech, and Language Processing.

[8]  Yanhui Guo,et al.  A Retinal Vessel Detection Approach Based on Shearlet Transform and Indeterminacy Filtering on Fundus Images , 2017, Symmetry.

[9]  Yanhui Guo,et al.  An effective clustering method based on data indeterminacy in neutrosophic set domain , 2018, Eng. Appl. Artif. Intell..

[10]  Florentin Smarandache,et al.  A unifying field in logics : neutrosophic logic : neutrosophy, neutrosophic set, neutrosophic probability , 2020 .

[11]  Abdolreza Rashno,et al.  Content-based image retrieval system with most relevant features among wavelet and color features , 2019, ArXiv.

[12]  Ming Zhang,et al.  A neutrosophic approach to image segmentation based on watershed method , 2010, Signal Process..

[13]  Behrouz Minaei-Bidgoli,et al.  Certainty of outlier and boundary points processing in data mining , 2018, 2019 27th Iranian Conference on Electrical Engineering (ICEE).

[14]  Alan F. Murray,et al.  Confidence estimation methods for neural networks : a practical comparison , 2001, ESANN.

[15]  Yanqing Zhang,et al.  Interval Neutrosophic Sets and Logic: Theory and Applications in Computing , 2005, ArXiv.

[16]  Behzad Nazari,et al.  Fully-automated segmentation of fluid regions in exudative age-related macular degeneration subjects: Kernel graph cut in neutrosophic domain , 2017, PloS one.

[17]  Keshab K. Parhi,et al.  Automated intra-retinal, sub-retinal and sub-RPE cyst regions segmentation in age-related macular degeneration (AMD) subjects , 2017 .

[18]  Yanhui Guo,et al.  NS-k-NN: Neutrosophic Set-Based k-Nearest Neighbors Classifier , 2017, Symmetry.

[19]  K Doi,et al.  An improved shift-invariant artificial neural network for computerized detection of clustered microcalcifications in digital mammograms. , 1996, Medical physics.

[20]  Tomohiro Nakatani,et al.  Noise robust speech recognition using recent developments in neural networks for computer vision , 2016, 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[21]  Keshab K. Parhi,et al.  Automated Fluid/Cyst Segmentation: A Quantitative Assessment of Diabetic Macular Edema , 2017 .

[22]  Abed Heshmati,et al.  Scheme for unsupervised colour-texture image segmentation using neutrosophic set and non-subsampled contourlet transform , 2016, IET Image Process..

[23]  Tara N. Sainath,et al.  Deep Convolutional Neural Networks for Large-scale Speech Tasks , 2015, Neural Networks.

[24]  Malay Kumar Kundu,et al.  Accurate segmentation of complex document image using digital shearlet transform with neutrosophic set as uncertainty handling tool , 2017, Appl. Soft Comput..

[25]  Yanhui Guo,et al.  New neutrosophic approach to image segmentation , 2009, Pattern Recognit..

[26]  Abdulkadir Sengür,et al.  A novel image edge detection algorithm based on neutrosophic set , 2014, Comput. Electr. Eng..

[27]  Yanhui Guo,et al.  KNCM: Kernel Neutrosophic c-Means Clustering , 2017, Appl. Soft Comput..

[28]  Abdulkadir Sengür,et al.  NCM: Neutrosophic c-means clustering algorithm , 2015, Pattern Recognit..

[29]  Keshab K. Parhi,et al.  Correlation between initial vision and vision improvement with automatically calculated retinal cyst volume in treated DME after resolution , 2017 .

[30]  F. Smarandache.  A Unifying Field in Logics: Neutrosophic Logic. , 1999 .

[31]  Jun Ye,et al.  A novel image thresholding algorithm based on neutrosophic similarity score , 2014 .

[32]  Keshab K. Parhi,et al.  Fully Automated Segmentation of Fluid/Cyst Regions in Optical Coherence Tomography Images With Diabetic Macular Edema Using Neutrosophic Sets and Graph Algorithms , 2018, IEEE Transactions on Biomedical Engineering.

[33]  Philip C. Woodland,et al.  Very deep convolutional neural networks for robust speech recognition , 2016, 2016 IEEE Spoken Language Technology Workshop (SLT).

[34]  Abdolreza Rashno,et al.  Automatic Segmentation of Choroid Layer in EDI OCT Images Using Graph Theory in Neutrosophic Space , 2018, 1812.01989.

[35]  Keshab K. Parhi,et al.  OCT Fluid Segmentation using Graph Shortest Path and Convolutional Neural Network* , 2018, 2018 40th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC).

[36]  Robert LIN,et al.  NOTE ON FUZZY SETS , 2014 .

[37]  Babak Nasersharif,et al.  Multiresolution convolutional neural network for robust speech recognition , 2017, 2017 Iranian Conference on Electrical Engineering (ICEE).

[38]  Florentin Smarandache,et al.  Refined neutrosophic sets in content-based image retrieval application , 2017, 2017 10th Iranian Conference on Machine Vision and Image Processing (MVIP).

[39]  Baiqing Sun,et al.  A novel glomerular basement membrane segmentation using neutrsophic set and shearlet transform on microscopic images , 2017, Health Inf. Sci. Syst..

[40]  Mrityunjaya V. Latte,et al.  Combined endeavor of Neutrosophic Set and Chan-Vese model to extract accurate liver image from CT scan , 2017, Comput. Methods Programs Biomed..