Computationally Efficient Spatio-Temporal Dynamic Texture Recognition for Volatile Organic Compound (VOC) Leakage Detection in Industrial Plants

Badawi, Diaa; Pan, Hongyi; Cetin, Sinan Cem; Çetin, A. Enis

doi:10.1109/jstsp.2020.2976555

Cited by 12 publications

(3 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Chemical sensors are widely used to detect ammonia, methane, and other Volatile Organic Compounds (VOCs) [1]- [3]. The life and performance of chemical gas detection sensors can be affected by various factors, including temperature, humidity, other interfering chemical gases, physical factors etc.…”

Section: Introductionmentioning

confidence: 99%

Detecting Anomaly in Chemical Sensors via L1-Kernel-Based Principal Component Analysis

et al. 2022

Self Cite

View full text Add to dashboard Cite

We propose a kernel-PCA based method to detect anomaly in chemical sensors. We use temporal signals produced by chemical sensors to form vectors to perform the Principal Component Analysis (PCA). We estimate the kernel-covariance matrix of the sensor data and compute the eigenvector corresponding to the largest eigenvalue of the covariance matrix. The anomaly can be detected by comparing the difference between the actual sensor data and the reconstructed data from the dominant eigenvector. In this paper, we introduce a new multiplication-free kernel, which is related to the ℓ 1 -norm for the anomaly detection task. The ℓ 1 -kernel PCA is not only computationally efficient but also energy-efficient because it does not require any actual multiplications during the kernel covariance matrix computation. Our experimental results show that our kernel-PCA method achieves a higher area under curvature (AUC) score (0.7483) than the baseline regular PCA method (0.7366).

show abstract

Section: Introductionmentioning

confidence: 99%

Detecting Anomaly in Chemical Sensors via L1-Kernel-Based Principal Component Analysis

et al. 2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…Convolutional neural networks (CNNs) have produced remarkable results in image classification [1][2][3][4][5][6][7][8][9][10][11][12], object detection [13][14][15][16][17] and semantic segmentation [18][19][20][21][22]. One of the most widely used and successful CNNs is ResNet [6], which can build very deep networks.…”

Section: Introductionmentioning

confidence: 99%

DCT Perceptron Layer: A Transform Domain Approach for Convolution Layer

Hongyi¹,

Zhu²,

Salih³

et al. 2022

Preprint

View full text Add to dashboard Cite

In this paper, we propose a novel Discrete Cosine Transform (DCT)-based neural network layer which we call DCT-perceptron to replace the 3 × 3 Conv2D layers in the Residual neural Network (ResNet). Convolutional filtering operations are performed in the DCT domain using element-wise multiplications by taking advantage of the Fourier and DCT Convolution theorems. A trainable softthresholding layer is used as the nonlinearity in the DCT perceptron. Compared to ResNet's Conv2D layer which is spatial-agnostic and channel-specific, the proposed layer is location-specific and channel-specific. The DCT-perceptron layer reduces the number of parameters and multiplications significantly while maintaining comparable accuracy results of regular ResNets in CIFAR-10 and ImageNet-1K. Moreover, the DCT-perceptron layer can be inserted with a batch normalization layer before the global average pooling layer in the conventional ResNets as an additional layer to improve classification accuracy.

show abstract

“…Deep convolution neural networks (CNN) are universally used in a wide range of applications including image classification [1,2,3,4,5,6,7,8,9], object detection [10,11,12,13] and semantic segmentation [14,15,16,17,18]. On the other hand, implementing deep neural networks in real-time resource-constrained environments such as embedded devices is very difficult due to insufficient memory and limited computational capacity.…”

Section: Introductionmentioning

confidence: 99%

Fast Walsh-Hadamard Transform and Smooth-Thresholding Based Binary Layers in Deep Neural Networks

Pan¹,

Dabawi²,

Çetin³

2021

Preprint

Self Cite

View full text Add to dashboard Cite

In this paper, we propose a novel layer based on fast Walsh-Hadamard transform (WHT) and smooththresholding to replace 1 × 1 convolution layers in deep neural networks.In the WHT domain, we denoise the transform domain coefficients using the new smooththresholding non-linearity, a smoothed version of the wellknown soft-thresholding operator. We also introduce a family of multiplication-free operators from the basic 2×2 Hadamard transform to implement 3 × 3 depthwise separable convolution layers. Using these two types of layers, we replace the bottleneck layers in MobileNet-V2 to reduce the network's number of parameters with a slight loss in accuracy. For example, by replacing the final third bottleneck layers, we reduce the number of parameters from 2.270M to 947K. This reduces the accuracy from 95.21% to 92.88% on the CIFAR-10 dataset. Our approach significantly improves the speed of data processing. The fast Walsh-Hadamard transform has a computational complexity of O(m log 2 m). As a result, it is computationally more efficient than the 1 × 1 convolution layer. The fast Walsh-Hadamard layer processes a tensor in R 10×32×32×1024 about 2 times faster than 1 × 1 convolution layer on NVIDIA Jetson Nano computer board.

show abstract

Computationally Efficient Spatio-Temporal Dynamic Texture Recognition for Volatile Organic Compound (VOC) Leakage Detection in Industrial Plants

Cited by 12 publications

References 20 publications

Detecting Anomaly in Chemical Sensors via L1-Kernel-Based Principal Component Analysis

Detecting Anomaly in Chemical Sensors via L1-Kernel-Based Principal Component Analysis

DCT Perceptron Layer: A Transform Domain Approach for Convolution Layer

Fast Walsh-Hadamard Transform and Smooth-Thresholding Based Binary Layers in Deep Neural Networks

Contact Info

Product

Resources

About