China 
Africa 
Explore chapters and articles related to this topic
Standards and Protocols for Agro-IoT
Published in Saravanan Krishnan, J Bruce Ralphin Rose, N R Rajalakshmi, Narayanan Prasanth, Cloud IoT Systems for Smart Agricultural Engineering, 2022
S. Mythili, K. Nithya, M. Krishnamoorthi, M. Kalamani
After filtering the region of interest from the image has to be extracted through a technique known as segmentation. In segmentation, the entire image is divided into many portions, which leads to analyzing the required image in a clear way, which thus helps create the exact map of the analyzing part. The simplest thresholding method is used where the grayscale image is converted into a binary scale image and based on the average value the threshold is set and the analysis is done. The next process is a feature extraction where the desired attributes of the data are identified and driven with the non-redundant data for further process. The enhanced predictability is achieved through Gabor filter and Gray Level Co-occurrence Matrix (GLCM) statistical features. The Gabor filter is a bandpass filter in which it provides the frequency and orientation representations of an image. Gabor features are best and outperforms when compared to GLCM. The distinctive features are classified by using classification algorithms such as KNN, Naive Bayes, Random Forest, Decision tree, and SVM.
Information Fusion for Multimodal Analysis and Recognition
Published in Ling Guan, Yifeng He, Sun-Yuan Kung, Multimedia Image and Video Processing, 2012
Yongjin Wang, Ling Guan, Anastasios N. Venetsanopoulos
The resulting facial image is normalized to a size of 64 × 64. A Gabor filter bank of five scales and eight orientations is then applied for feature extraction. Using Gabor wavelet features to represent facial expression has been explored and shown to be very efficient in the literature [36]. It allows a description of spatial frequency structure in the image while preserving information about spatial relations. In this work, the Gabor filter bank is implemented using the algorithm proposed in [37]. Due to the large dimensionality of the Gabor coefficients, we downsample each subband to a size of 32 × 32, and then perform dimensionality reduction on all the down-sampled Gabor coefficients using the PCA method.
Feature Analysis and Pattern Classification
Published in Scott E. Umbaugh, Digital Image Processing and Analysis, 2017
The Gabor filter can be viewed as a sinusoidal wave of specified frequency and orientation, convolved by a Gaussian envelope. A 2D Gabor filter acts as a local band-pass filter with specific frequency and orientation. Mathematically, in the spatial domain, a 2D Gabor filter is the result of multiplication of a 2D Gaussian function and a complex exponential function, which can be represented as follows: g(λ,θ,φ,σ,γ)=exp(−c(θ)2+γ2r(θ)22σ2)exp(j(2πc(θ)λ+φ)) where λ is the wavelength, so 1/λ is the frequency, θ is the orientation angle, φ is the phase offset, σ represents the standard deviation of the Gaussian factor, γ is the aspect ratio of the Gabor equation, r is the row, and c is the column coordinate.
Quantitative prediction of fracture toughness (K Ic ) of polymer by fractography using deep neural networks
Published in Science and Technology of Advanced Materials: Methods, 2022
Y. Mototake, K. Ito, M. Demura
The DNN model used in this study was the vgg16 model [18], which identifies natural images such as animals and vehicles. vgg16 is composed of convolution, pooling, and fully connected layers. A convolution layer convolves the input space using a convolution filter. Convolution filters of DNNs trained on natural images frequently behave similarly to Gabor filters. A Gabor filter can extract image features such as the edges of graphic structures in images. In vgg16, a structure comprising a stack of convolution layers alternated with pooling layers is formed, and the pooling layers compress the dimensions by replacing certain regions of the convolution layers with representative values such as the maximum and average of the region. Pooling can be viewed as a transformation that renders the input space coarse grained. Thus, after training on natural images, vgg16 is expected to extract useful features on various scales for natural images. In the prediction of using fracture surface images, the evaluation of the edges of the graphic structures in the images is also considered important. This could be the reason why the proposed framework worked effectively.
Image segmentation using K-means clustering, Gabor filter and moving mesh method
Published in The Imaging Science Journal, 2021
In the task of tumour segmentation, the large difference in shape and texture between different tumours increases the difficulty of medical image processing. Gabor filter is thought to be a good model for recognizing texture and has better noise robustness. The Gabor filter was proposed by Daugman [7] and uses complex functions as the basis of Fourier transformations in applications of information theory. An important feature of Gabor wavelet is that the product of its standard deviation is minimal in both the time domain and frequency domain. Gabor filter can be used to extract feature information from images and is widely used in texture segmentation [8] and human identification [9]. Haghighat et al. [10] recommended to perform biometric verification in a cloud environment. Kumar et al. [11] proposed a new unsupervised retinal blood vessel segmentation method. In the case of uneven illumination, Gabor filter can effectively decrease the noise of the image and distinguish the directional structures such as cracks. Also, Gabor filter possesses optimal localization properties in both the spatial and frequency domain, so it can well describe the local structure information corresponding to the spatial scale and orientation selectivity. However, traditional Gabor filter is difficult to extract rotation-invariant texture features. Therefore, this study modifies the traditional Gabor filter into a circular version.
Facial Action Unit Intensity Detection by Extracting Complimentary Information using Distance Metric Learning
Published in IETE Journal of Research, 2020
Neeru Rathee, Dinesh Ganotra, Ajay Rathee
Gabor wavelet is one of the most famous techniques for representing texture information of an object. A Gabor filter is defined with a Gaussian kernel that is modulated with a sinusoidal plane. To extract the texture information, filter bank consisting of Gabor filters with various scales and rotations is created. Gabor features extract information about the small region around the facial feature key point. A detail of facial feature extraction using Gabor filters is presented by Tian et al. [14]. Recently, Li et al. [3] have applied Gabor features around landmark points for continuous FACS intensity estimation. In the proposed approach, we have implemented 40 Gabor filters (5 scales and 8 orientations) on the facial images so as to extract a feature vector of size 2820.