Enhancement of Deep Learning in Image Classification Performance Using Xception with the Swish Activation Function for Colorectal Polyp Preliminary Screening

Jinsakul, Natinai; Tsai, Cheng-Fa; Tsai, Chuen Jinn; Wu, Pensée

doi:10.3390/math7121170

Cited by 43 publications

(24 citation statements)

References 51 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is a deep learning image classifier using depthwise separable convolution layers with residual connections, which has been pre-trained on large scale images [ 26 ]. After input, data using only pointwise convolution (1 × 1 convolution) create separate convolution sizes of 3 × 3 without average pooling, which proceeds in nonoverlapping sections of the output channels to then be fed-forward for concatenation [ 26 , 27 ]. The model demonstrates a strong ability to generalize to images outside the original dataset via transfer learning, such as feature extraction and fine-tuning.…”

Section: Processing Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Silent Speech Decoding Using Spectrogram Features Based on Neuromuscular Activities

Wang

Zhang

et al. 2020

Brain Sciences

View full text Add to dashboard Cite

Silent speech decoding is a novel application of the Brain–Computer Interface (BCI) based on articulatory neuromuscular activities, reducing difficulties in data acquirement and processing. In this paper, spatial features and decoders that can be used to recognize the neuromuscular signals are investigated. Surface electromyography (sEMG) data are recorded from human subjects in mimed speech situations. Specifically, we propose to utilize transfer learning and deep learning methods by transforming the sEMG data into spectrograms that contain abundant information in time and frequency domains and are regarded as channel-interactive. For transfer learning, a pre-trained model of Xception on the large image dataset is used for feature generation. Three deep learning methods, Multi-Layer Perception, Convolutional Neural Network and bidirectional Long Short-Term Memory, are then trained using the extracted features and evaluated for recognizing the articulatory muscles’ movements in our word set. The proposed decoders successfully recognized the silent speech and bidirectional Long Short-Term Memory achieved the best accuracy of 90%, outperforming the other two algorithms. Experimental results demonstrate the validity of spectrogram features and deep learning algorithms.

show abstract

Section: Processing Methodsmentioning

confidence: 99%

“…So, synergistic features from multi-channel sEMG are considered to recognize different words. Xception, originally designed for image classification [ 26 , 27 , 28 , 29 ], is utilized to process spectrograms of multichannel sEMG to explore the spatial correlation.…”

Section: Introductionmentioning

confidence: 99%

Silent Speech Decoding Using Spectrogram Features Based on Neuromuscular Activities

Wang

Zhang

et al. 2020

Brain Sciences

View full text Add to dashboard Cite

show abstract

“…Xception [11] is meant to be a hypothesis reliant Inception module used in developing correlations of crosschannel as well as spatial relations inside feature maps of CNN that is capable of isolating the model. The typical Inception module from Inception v3, a module which has employed cross-channel relations by isolating the input data in 4 phases for convolution size of 1 x 1, average pooling, maps correlations of convolution size 3 x 3 and send them for combination.…”

Section: Dlxm Based Feature Extractionmentioning

confidence: 99%

Improvement in Tongue Color Image Analysis for Disease Identification Using Deep Learning Based Depthwise Separable Convolution Model

Rajakumaran¹,

Sasikala²

2021

INDJCSE

View full text Add to dashboard Cite

Disease diagnosis using tongue color image is a traditional non-invasive method widely employed to determine the status of the patient's internal organ. The elimination of dependencies on subjective and expert knowledge assessment for tongue diagnosis might considerably raise the scope of wide utilization of tongue diagnosis over the globe, including Western medicine. Computer based tongue diagnosis connected to light estimation, color correction, tongue segmentation, image analysis, geometry analysis, etc is a proficient method to diagnose diseases. This paper introduces a new Deep Learning with Depthwise Separable Convolution (Xception) Model called DLXM for tongue color image analysis. The presented model involves data augmentation and bilateral filtering (BF) based noise removal at the preprocessing stage. Additionally, the DLXM is applied for feature extraction process. At last, the bagging classifier (BC)and multilayer perceptron classifier (MLPC) models are employed to categorize the feature vectors into distinct types of diseases. The performance of the presented model is evaluated against benchmark tongue image dataset and the results depicted the effectual classification performance on the applied images. The experimental values notified that the DLXM-MLPC model has outperformed the compared methods by achieving a higher precision, recall, accuracy, and F1-Score of 97.35%, 97.01%, 97.01%, and 96.77% respectively.

show abstract

“…Swish is superior in terms of expressive power and fast convergence. It has been applied in diverse deep learning applications and proved to be more effective than ReLU and other activation functions (Tripathi et al, 2019;Jinsakul et al, 2019;Wang et al, 2019).…”

Section: Predefined Activation Functionsmentioning

confidence: 99%

Parametric Flatten-T Swish: An Adaptive Nonlinear Activation Function for Deep Learning

Chieng

Wahid

Ong

2020

Journal of Information and Communication Technology

View full text Add to dashboard Cite

QActivation function is a key component in deep learning that performs non-linear mappings between the inputs and outputs. Rectified Linear Unit (ReLU) has been the most popular activation function across the deep learning community. However, ReLU contains several shortcomings that can result in inefficient training of the deep neural networks, these are: 1) the negative cancellation property of ReLU tends to treat negative inputs as unimportant information for the learning, resulting in performance degradation; 2) the inherent predefined nature of ReLU is unlikely to promote additional flexibility, expressivity, and robustness to the networks; 3) the mean activation of ReLU is highly positive and leads to bias shift effect in network layers; and 4) the multilinear structure of ReLU restricts the non-linear approximation power of the networks. To tackle these shortcomings, this paper introduced Parametric Flatten-T Swish (PFTS) as an alternative to ReLU. By taking ReLU as a baseline method, the experiments showed that PFTS improved classification accuracy on SVHN dataset by 0.31%, 0.98%, 2.16%, 17.72%, 1.35%, 0.97%, 39.99%, and 71.83% on DNN-3A, DNN-3B, DNN-4, DNN-5A, DNN-5B, DNN-5C, DNN-6, and DNN-7, respectively. Besides, PFTS also achieved the highest mean rank among the comparison methods. The proposed PFTS manifested higher non-linear approximation power during training and thereby improved the predictive performance of the networks.

show abstract

Enhancement of Deep Learning in Image Classification Performance Using Xception with the Swish Activation Function for Colorectal Polyp Preliminary Screening

Cited by 43 publications

References 51 publications

Silent Speech Decoding Using Spectrogram Features Based on Neuromuscular Activities

Silent Speech Decoding Using Spectrogram Features Based on Neuromuscular Activities

Improvement in Tongue Color Image Analysis for Disease Identification Using Deep Learning Based Depthwise Separable Convolution Model

Parametric Flatten-T Swish: An Adaptive Nonlinear Activation Function for Deep Learning

Contact Info

Product

Resources

About