Search results for: depthwise separable convolutions

Authors: Jianghao Hu, Hongyu Wang

Abstract:

The effective and efficient human pose estimation algorithm is an important task for real-time human pose estimation on mobile devices. This paper proposes a light-weight human key points detection algorithm, Light-Weight Network for Real-Time Pose Estimation (LWPE). LWPE uses light-weight backbone network and depthwise separable convolutions to reduce parameters and lower latency. LWPE uses the feature pyramid network (FPN) to fuse the high-resolution, semantically weak features with the low-resolution, semantically strong features. In the meantime, with multi-scale prediction, the predicted result by the low-resolution feature map is stacked to the adjacent higher-resolution feature map to intermediately monitor the network and continuously refine the results. At the last step, the key point coordinates predicted in the highest-resolution are used as the final output of the network. For the key-points that are difficult to predict, LWPE adopts the online hard key points mining strategy to focus on the key points that hard predicting. The proposed algorithm achieves excellent performance in the single-person dataset selected in the AI (artificial intelligence) challenge dataset. The algorithm maintains high-precision performance even though the model only contains 3.9M parameters, and it can run at 225 frames per second (FPS) on the generic graphics processing unit (GPU).

Keywords: depthwise separable convolutions, feature pyramid network, human pose estimation, light-weight backbone

Procedia PDF Downloads 122

Authors: Kamal KC, Zhendong Yin, Dasen Li, Zhilu Wu

Abstract:

Intra-plant disease classification based on leaf images is a challenging computer vision task due to similarities in texture, color, and shape of leaves with a slight variation of leaf spot; and external environmental changes such as lighting and background noises. Deep convolutional neural network (DCNN) has proven to be an effective tool for binary classification. In this paper, two methods for binary classification of diseased plant leaves using DCNN are presented; model created from scratch and transfer learning. Our main contribution is a thorough evaluation of 4 networks created from scratch and transfer learning of 5 pre-trained models. Training and testing of these models were performed on a plant leaf images dataset belonging to 16 distinct classes, containing a total of 22,265 images from 8 different plants, consisting of a pair of healthy and diseased leaves. We introduce a deep CNN model, Optimized MobileNet. This model with depthwise separable CNN as a building block attained an average test accuracy of 99.77%. We also present a fine-tuning method by introducing the concept of a convolutional block, which is a collection of different deep neural layers. Fine-tuned models proved to be efficient in terms of accuracy and computational cost. Fine-tuned MobileNet achieved an average test accuracy of 99.89% on 8 pairs of [healthy, diseased] leaf ImageSet.

Keywords: deep convolution neural network, depthwise separable convolution, fine-grained classification, MobileNet, plant disease, transfer learning

Procedia PDF Downloads 156

Authors: Newroz Nooralddin Abdulrazaq, Thuraya Mahmood Qaradaghi

Abstract:

The McEliece cryptosystem is an asymmetric type of cryptography based on error correction code. The classical McEliece used irreducible binary Goppa code which considered unbreakable until now especially with parameter [1024, 524, and 101], but it is suffering from large public key matrix which leads to be difficult to be used practically. In this work Irreducible and Separable Goppa codes have been introduced. The Irreducible and Separable Goppa codes used are with flexible parameters and dynamic error vectors. A Comparison between Separable and Irreducible Goppa code in McEliece Cryptosystem has been done. For encryption stage, to get better result for comparison, two types of testing have been chosen; in the first one the random message is constant while the parameters of Goppa code have been changed. But for the second test, the parameters of Goppa code are constant (m=8 and t=10) while the random message have been changed. The results show that the time needed to calculate parity check matrix in separable are higher than the one for irreducible McEliece cryptosystem, which is considered expected results due to calculate extra parity check matrix in decryption process for g2(z) in separable type, and the time needed to execute error locator in decryption stage in separable type is better than the time needed to calculate it in irreducible type. The proposed implementation has been done by Visual studio C#.

Keywords: McEliece cryptosystem, Goppa code, separable, irreducible

Procedia PDF Downloads 236

Authors: Jiangnan Yang, Shuangli Liu, Jinjun Wu, Xinyu Su, Nan Hai

Abstract:

These recent years have witnessed that convolutional neural network (CNN)-based methods for detecting infrared dim small targets have achieved outstanding performance. However, these methods typically employ standard convolutions, neglecting to fully consider the spatial characteristics of the pixel distribution of infrared dim small targets. Therefore, based on interleaved group convolutions, we propose a Pinwheel-shaped convolution (PConv) to replace the first two layers of standard convolutions in the backbone network. Compared to standard convolutions, PConv better conforms to the pixel spatial distribution of dim small targets and introduces only a minimal increase in parameters while effectively enhancing the extraction of features from dim small targets. Additionally, the latest loss function, combining intersection over union (IoU) loss and distance loss, fails to adequately account for the sensitivity differences in scale and position for targets of different sizes, thereby underutilizing the detection performance of neural networks for weak, small targets. To address this, we introduce dynamic completely IoU (DCIoU) loss, dynamically adjusting the influence coefficients of IoU and distance based on the size of the target bounding box, enhancing the neural network's ability to converge on targets of varying scales, thereby significantly boosting detection accuracy. Finally, we have curated the largest and most challenging dataset of real-shot single-frame infrared dim small target detection to date: SIRST-UAVB. Integrating PConv and DCIoU into the state-of-the-art small target detection algorithm, we conduct tests on both public datasets and SIRST-UAVB, achieving significant performance improvements, thus validating the effectiveness and generalizability of our approach.

Keywords: infrared dim small target detection, deep learning, interleaved group convolutions, feature extraction, loss function, dataset

Procedia PDF Downloads 5

Authors: Ananya Ananya, Karthik Rao

Abstract:

Accurate segmentation of knee cartilage in 3-D magnetic resonance (MR) images for quantitative assessment of volume is crucial for studying and diagnosing osteoarthritis (OA) of the knee, one of the major causes of disability in elderly people. Radiologists generally perform this task in slice-by-slice manner taking 15-20 minutes per 3D image, and lead to high inter and intra observer variability. Hence automatic methods for knee cartilage segmentation are desirable and are an active field of research. This paper presents design and experimental evaluation of 2D convolutional neural networks based fully automated methods for knee cartilage segmentation in 3D MRI. The architectures are validated based on 40 test images and 60 training images from SKI10 dataset. The proposed methods segment 2D slices one by one, which are then combined to give segmentation for whole 3D images. Proposed methods are modified versions of U-net and dilated convolutions, consisting of a single step that segments the given image to 5 labels: background, femoral cartilage, tibia cartilage, femoral bone and tibia bone; cartilages being the primary components of interest. U-net consists of a contracting path and an expanding path, to capture context and localization respectively. Dilated convolutions lead to an exponential expansion of receptive field with only a linear increase in a number of parameters. A combination of modified U-net and dilated convolutions has also been explored. These architectures segment one 3D image in 8 – 10 seconds giving average volumetric Dice Score Coefficients (DSC) of 0.950 - 0.962 for femoral cartilage and 0.951 - 0.966 for tibia cartilage, reference being the manual segmentation.

Keywords: convolutional neural networks, dilated convolutions, 3 dimensional, fully automated, knee cartilage, MRI, segmentation, U-net

Procedia PDF Downloads 231

Authors: Qinyi Mei, Li-Ping Wang

Abstract:

MDS matrices are of great significance in the design of block ciphers and hash functions. In the present paper, we investigate the problem of constructing MDS matrices which are both lightweight and low-latency. We propose a new method of constructing lightweight MDS matrices using circulant matrices which can be implemented efficiently in hardware. Furthermore, we provide circulant MDS matrices with as few bit XOR operations as possible for the classical dimensions 4 × 4, 8 × 8 over the space of linear transformations over finite field F42 . In contrast to previous constructions of MDS matrices, our constructions have achieved fewer XORs.

Keywords: linear diffusion layer, circulant matrix, lightweight, maximum distance separable (MDS) matrix

Procedia PDF Downloads 380

Authors: Kaustav Mukherjee

Abstract:

In this paper we propose a Neural Style Transfer solution whereby we have created a Lightweight Separable Convolution Kernel Based GAN Architecture (SC-GAN) which will very useful for designing filter for Mobile Phone Cameras and also Edge Devices which will convert any image to its 2D ANIMATED COMIC STYLE Movies like HEMAN, SUPERMAN, JUNGLE-BOOK. This will help the 2D animation artist by relieving to create new characters from real life person's images without having to go for endless hours of manual labour drawing each and every pose of a cartoon. It can even be used to create scenes from real life images.This will reduce a huge amount of turn around time to make 2D animated movies and decrease cost in terms of manpower and time. In addition to that being extreme light-weight it can be used as camera filters capable of taking Comic Style Shots using mobile phone camera or edge device cameras like Raspberry Pi 4,NVIDIA Jetson NANO etc. Existing Methods like CartoonGAN with the model size close to 170 MB is too heavy weight for mobile phones and edge devices due to their scarcity in resources. Compared to the current state of the art our proposed method which has a total model size of 31 MB which clearly makes it ideal and ultra-efficient for designing of camera filters on low resource devices like mobile phones, tablets and edge devices running OS or RTOS. .Owing to use of high resolution input and usage of bigger convolution kernel size it produces richer resolution Comic-Style Pictures implementation with 6 times lesser number of parameters and with just 25 extra epoch trained on a dataset of less than 1000 which breaks the myth that all GAN need mammoth amount of data. Our network reduces the density of the Gan architecture by using Depthwise Separable Convolution which does the convolution operation on each of the RGB channels separately then we use a Point-Wise Convolution to bring back the network into required channel number using 1 by 1 kernel.This reduces the number of parameters substantially and makes it extreme light-weight and suitable for mobile phones and edge devices. The architecture mentioned in the present paper make use of Parameterised Batch Normalization Goodfellow etc al. (Deep Learning OPTIMIZATION FOR TRAINING DEEP MODELS page 320) which makes the network to use the advantage of Batch Norm for easier training while maintaining the non-linear feature capture by inducing the learnable parameters

Keywords: comic stylisation from camera image using GAN, creating 2D animated movie style custom stickers from images, depth-wise separable convolutional neural network for light-weight GAN architecture for EDGE devices, GAN architecture for 2D animated cartoonizing neural style, neural style transfer for edge, model distilation, perceptual loss

Procedia PDF Downloads 97

Authors: Kun Xu, Yuan Xu, Jia Qiao

Abstract:

The document detection plays an important role in optical character recognition and text analysis. Because the traditional detection methods have weak generalization ability, and deep neural network has complex structure and large number of parameters, which cannot be well applied in mobile devices, this paper proposes a lightweight Document Corner Detection Network (DCDNet). DCDNet is a two-stage architecture. The first stage with Encoder-Decoder structure adopts depthwise separable convolution to greatly reduce the network parameters. After introducing the Feature Attention Union (FAU) module, the second stage enhances the feature information of spatial and channel dim and adaptively adjusts the size of receptive field to enhance the feature expression ability of the model. Aiming at solving the problem of the large difference in the number of pixel distribution between corner and non-corner, Weighted Binary Cross Entropy Loss (WBCE Loss) is proposed to define corner detection problem as a classification problem to make the training process more efficient. In order to make up for the lack of Dataset of document corner detection, a Dataset containing 6620 images named Document Corner Detection Dataset (DCDD) is made. Experimental results show that the proposed method can obtain fast, stable and accurate detection results on DCDD.

Keywords: document detection, corner detection, attention mechanism, lightweight

Procedia PDF Downloads 324

Authors: Jamal Laaouine, Mohammed Elhassani Charkani

Abstract:

Let p be a prime and let b be an integer. MDS b-symbol codes are a direct generalization of MDS codes. The γ-constacyclic codes of length pˢ over the finite commutative chain ring Fₚm [u]/ < u² > had been classified into four distinct types, where is a nonzero element of the field Fₚm. Let C₃ be a code of Type 3. In this paper, we obtain the b-symbol distance db(C₃) of the code C₃. Using this result, necessary and sufficient conditions under which C₃ is an MDS b-symbol code are given.

Keywords: constacyclic code, repeated-root code, maximum distance separable, MDS codes, b-symbol distance, finite chain rings

Procedia PDF Downloads 106

Authors: Nuray Güy, Mahmut Özacar

Abstract:

In the past few decades, considerable attention has been devoted to the photocatalysts for the photocatalytic degradation of environmental pollutants. Many novel nanostructured photocatalysts for wastewater treatment have been investigated, such as TiO2 and, CdS, ZnO and silver halides (AgX, X = Cl, Br, I). The silver halides are photosensitive materials which can absorb photons in the visible region to produce electron–hole pairs. Silver halides are expensive that restricts their applications in large-scale photocatalytic processes. Tannin contains hydroxyl functional groups, it was employed as a modifier to improve the surface properties and adsorption capacity of the activated carbon towards the metal cations uptake. In this work, we designed a new structure of magnetically separable photocatalyst that combines AgCl/ZnO nanoparticles with Fe3O4 nanoparticles deposited on tannin, which was denoted as (AgI/ZnO)-Fe3O4/Tannin. The as-prepared products are characterized by X-ray diffraction (XRD), field emission scanning electron microscope (FESEM), Fourier transform infrared (FTIR), diffuse reflectance spectra (DRS) and vibrating sample magnetometer (VSM). The photocatalyst exhibited high activity degrading a textile dye under visible light irradiation. Moreover, the excellent magnetic property gives a more convenient way to recycle the photocatalysts.

Keywords: AgI/ZnO-Fe3O4/Tannin, visible light, magnetically separable, photocatalyst

Procedia PDF Downloads 193

Authors: Bin Liu, Weijie Liu, Bin Sun, Yihui Luo

Abstract:

In order to solve the problem of lower spatial resolution and block effect in the fusion method based on separable wavelet transform in the resulting fusion image, a new sampling mode based on multi-resolution analysis of two-channel non separable wavelet transform, whose dilation matrix is [1,1;1,-1], is presented and a multispectral image fusion method based on this kind of sampling mode is proposed. Filter banks related to this kind of wavelet are constructed, and multiresolution decomposition of the intensity of the MS and panchromatic image are performed in the sampled mode using the constructed filter bank. The low- and high-frequency coefficients are fused by different fusion rules. The experiment results show that this method has good visual effect. The fusion performance has been noted to outperform the IHS fusion method, as well as, the fusion methods based on DWT, IHS-DWT, IHS-Contourlet transform, and IHS-Curvelet transform in preserving both spectral quality and high spatial resolution information. Furthermore, when compared with the fusion method based on nonsubsampled two-channel non separable wavelet, the proposed method has been observed to have higher spatial resolution and good global spectral information.

Keywords: image fusion, two-channel sampled nonseparable wavelets, multispectral image, panchromatic image

Procedia PDF Downloads 403

Authors: A. Maghari, V. M. Maleki

Abstract:

In this paper, we present a quantum statistical mechanical formulation from our recently analytical expressions for partial-wave transition matrix of a three-particle system. We report the quantum reactive cross sections for three-body scattering processes 1 + (2,3)-> 1 + (2,3) as well as recombination 1 + (2,3) -> 2 + (3,1) between one atom and a weakly-bound dimer. The analytical expressions of three-particle transition matrices and their corresponding cross-sections were obtained from the three-dimensional Faddeev equations subjected to the rank-two non-local separable potentials of the generalized Yamaguchi form. The equilibrium quantum statistical mechanical properties such partition function and equation of state as well as non-equilibrium quantum statistical properties such as transport cross-sections and their corresponding transport collision integrals were formulated analytically. This leads to obtain the transport properties, such as viscosity and diffusion coefficient of a moderate dense gas.

Keywords: statistical mechanics, nonlocal separable potential, three-body interaction, faddeev equations

Procedia PDF Downloads 374

Authors: Priya Arora, Jaspreet K. Rajput

Abstract:

Superparamagnetism has accelerated the research and use of more economical and ecological magnetically separable catalysts due to their more efficient isolation by response to an applied magnetic field. Magnetite nanomaterials coated by SiO2 shell have received a great deal of focus in the last decades as it provides high stability and also easy further surface functionalization depending upon the application. In this protocol, Fe3O4 magnetic nanoparticles have been synthesized by co-precipitation combined with sonication method. Further, Fe3O4 superparamagnetic nanoparticles have been functionalized by various moieties to serve as efficient catalysts for multicomponent reactions. The functionalized nanoparticles were characterized by techniques like Fourier transform infrared spectroscopy (FT-IR), transmission electron microscopy (TEM), scanning electron microscopy (SEM), vibrating sample magnetometer (VSM), thermogravimetric analysis (TGA), Brunauer-Emmett-Teller (BET) surface area analysis. The as prepared nanocatalysts can be reused for several times without any significant loss in its activity. The utilization of magnetic nanoparticles as catalysts for this reaction is one approach i.e. green, inexpensive, facile and widely applicable.

Keywords: functionalized, magnetite, multicomponent reactions, superparamagnetic

Procedia PDF Downloads 315

Authors: Kemal Polat

Abstract:

In this paper, an attribute weighting method called fuzzy C-means clustering based attribute weighting (FCMAW) for classification of Diabetes disease dataset has been used. The aims of this study are to reduce the variance within attributes of diabetes dataset and to improve the classification accuracy of classifier algorithm transforming from non-linear separable datasets to linearly separable datasets. Pima Indians Diabetes dataset has two classes including normal subjects (500 instances) and diabetes subjects (268 instances). Fuzzy C-means clustering is an improved version of K-means clustering method and is one of most used clustering methods in data mining and machine learning applications. In this study, as the first stage, fuzzy C-means clustering process has been used for finding the centers of attributes in Pima Indians diabetes dataset and then weighted the dataset according to the ratios of the means of attributes to centers of theirs. Secondly, after weighting process, the classifier algorithms including support vector machine (SVM) and k-NN (k- nearest neighbor) classifiers have been used for classifying weighted Pima Indians diabetes dataset. Experimental results show that the proposed attribute weighting method (FCMAW) has obtained very promising results in the classification of Pima Indians diabetes dataset.

Keywords: fuzzy C-means clustering, fuzzy C-means clustering based attribute weighting, Pima Indians diabetes, SVM

Procedia PDF Downloads 382

Authors: Jaeyoung Lee

Abstract:

Autonomous driving systems require high reliability to provide people with a safe and comfortable driving experience. However, despite the development of a number of vehicle sensors, it is difficult to always provide high perceived performance in driving environments that vary from time to season. The image segmentation method using deep learning, which has recently evolved rapidly, provides high recognition performance in various road environments stably. However, since the system controls a vehicle in real time, a highly complex deep learning network cannot be used due to time and memory constraints. Moreover, efficient networks are optimized for GPU environments, which degrade performance in embedded processor environments equipped simple hardware accelerators. In this paper, a semantic segmentation network, matrix multiplication accelerator network (MMANet), optimized for matrix multiplication accelerator (MMA) on Texas instrument digital signal processors (TI DSP) is proposed to improve the recognition performance of autonomous driving system. The proposed method is designed to maximize the number of layers that can be performed in a limited time to provide reliable driving environment information in real time. First, the number of channels in the activation map is fixed to fit the structure of MMA. By increasing the number of parallel branches, the lack of information caused by fixing the number of channels is resolved. Second, an efficient convolution is selected depending on the size of the activation. Since MMA is a fixed, it may be more efficient for normal convolution than depthwise separable convolution depending on memory access overhead. Thus, a convolution type is decided according to output stride to increase network depth. In addition, memory access time is minimized by processing operations only in L3 cache. Lastly, reliable contexts are extracted using the extended atrous spatial pyramid pooling (ASPP). The suggested method gets stable features from an extended path by increasing the kernel size and accessing consecutive data. In addition, it consists of two ASPPs to obtain high quality contexts using the restored shape without global average pooling paths since the layer uses MMA as a simple adder. To verify the proposed method, an experiment is conducted using perfsim, a timing simulator, and the Cityscapes validation sets. The proposed network can process an image with 640 x 480 resolution for 6.67 ms, so six cameras can be used to identify the surroundings of the vehicle as 20 frame per second (FPS). In addition, it achieves 73.1% mean intersection over union (mIoU) which is the highest recognition rate among embedded networks on the Cityscapes validation set.

Keywords: edge network, embedded network, MMA, matrix multiplication accelerator, semantic segmentation network

Procedia PDF Downloads 96

Authors: Jaskarn Singh Bhullar, Divya Taneja, Manish Gupta, Rajesh Kumar Narula

Abstract:

The existence conditions of dual containing constacyclic codes have opened a new path for finding quantum maximum distance separable (MDS) codes. Using these conditions parameters of length n=(q²+1)/2 quantum MDS codes were improved. A class of quantum MDS codes of length n=(q²+q+1)/h, where h>1 is an odd prime, have also been constructed having large minimum distance and these codes are new in the sense as these are not available in the literature.

Keywords: hermitian construction, constacyclic codes, cyclotomic cosets, quantum MDS codes, singleton bound

Procedia PDF Downloads 354

Authors: Nouha Khediri, Mohammed Ben Ammar, Monji Kherallah

Abstract:

Recently, facial emotion recognition (FER) has become increasingly essential to understand the state of the human mind. Accurately classifying emotion from the face is a challenging task. In this paper, we present a facial emotion recognition approach named CV-FER, benefiting from deep learning, especially CNN and VGG16. First, the data is pre-processed with data cleaning and data rotation. Then, we augment the data and proceed to our FER model, which contains five convolutions layers and five pooling layers. Finally, a softmax classifier is used in the output layer to recognize emotions. Based on the above contents, this paper reviews the works of facial emotion recognition based on deep learning. Experiments show that our model outperforms the other methods using the same FER2013 database and yields a recognition rate of 92%. We also put forward some suggestions for future work.

Keywords: CNN, deep-learning, facial emotion recognition, machine learning

Procedia PDF Downloads 61

Authors: Thibaud Berthomier, Ali Mansour, Luc Bressollette, Frédéric Le Roy, Dominique Mottier, Léo Fréchier, Barthélémy Hermenault

Abstract:

Deep Venous Thrombosis (DVT) occurs when a thrombus is formed within a deep vein (most often in the legs). This disease can be deadly if a part or the whole thrombus reaches the lung and causes a Pulmonary Embolism (PE). This disorder, often asymptomatic, has multifactorial causes: immobilization, surgery, pregnancy, age, cancers, and genetic variations. Our project aims to relate the thrombus epidemiology (origins, patient predispositions, PE) to its structure using ultrasound images. Ultrasonography and elastography were collected using Toshiba Aplio 500 at Brest Hospital. This manuscript compares two classification approaches: spectral clustering and scattering operator. The former is based on the graph and matrix theories while the latter cascades wavelet convolutions with nonlinear modulus and averaging operators.

Keywords: deep venous thrombosis, ultrasonography, elastography, scattering operator, wavelet, spectral clustering

Procedia PDF Downloads 450

Authors: Rogelio Luck, Yucheng Liu

Abstract:

This paper proposes the continuous-time singular value decomposition (SVD) for the impulse response function, a special kind of Green’s functions e⁻⁽ᵗ⁻ ᵀ⁾, in order to find a set of singular functions and singular values so that the convolutions of such function with the set of singular functions on a specified domain are the solutions to the inhomogeneous differential equations for those singular functions. A numerical example was illustrated to verify the proposed method. Besides the continuous-time SVD, a discrete-time SVD is also presented for the impulse response function, which is modeled using a Toeplitz matrix in the discrete system. The proposed method has broad applications in signal processing, dynamic system analysis, acoustic analysis, thermal analysis, as well as macroeconomic modeling.

Keywords: singular value decomposition, impulse response function, Green’s function , Toeplitz matrix , Hankel matrix

Procedia PDF Downloads 123

Authors: Kemal Polat

Abstract:

In this study, it has been estimated whether there is skin by using RBG values obtained from the camera and k-nearest neighbor (k-NN) classifier. The dataset used in this study has an unbalanced distribution and a linearly non-separable structure. This problem can also be called a big data problem. The Skin dataset was taken from UCI machine learning repository. As the classifier, we have used the k-NN method to handle this big data problem. For k value of k-NN classifier, we have used as 1. To train and test the k-NN classifier, 50-50% training-testing partition has been used. As the performance metrics, TP rate, FP Rate, Precision, recall, f-measure and AUC values have been used to evaluate the performance of k-NN classifier. These obtained results are as follows: 0.999, 0.001, 0.999, 0.999, 0.999, and 1,00. As can be seen from the obtained results, this proposed method could be used to predict whether the image is skin or not.

Keywords: k-NN classifier, skin or non-skin classification, RGB values, classification

Procedia PDF Downloads 219

Authors: Tian Lan, Yixiang Wang, Wenxin Tai, Yilan Lyu, Zufeng Wu

Abstract:

The monaural speech enhancement task in the time-frequencydomain has a myriad of approaches, with the stacked con-volutional neural network (CNN) demonstrating superiorability in feature extraction and selection. However, usingstacked single convolutions method limits feature represen-tation capability and generalization ability. In order to solvethe aforementioned problem, we propose an attention-basedadaptive convolutional network that integrates the multi-scale convolutional operations into a operation-specific blockvia input dependent attention to adapt to complex auditoryscenes. In addition, we introduce a two-stage progressivelearning method to enlarge the receptive field without a dra-matic increase in computation burden. We conduct a series ofexperiments based on the TIMIT corpus, and the experimen-tal results prove that our proposed model is better than thestate-of-art models on all metrics.

Keywords: speech enhancement, adaptive convolu-tion, progressive learning, time-frequency domain

Procedia PDF Downloads 89

Authors: Behnam Tavakkol

Abstract:

Uncertain data mining algorithms use different ways to consider uncertainty in data such as by representing a data object as a sample of points or a probability distribution. Fuzzy methods have long been used for clustering traditional (certain) data objects. They are used to produce non-crisp cluster labels. For uncertain data, however, besides some uncertain fuzzy k-medoids algorithms, not many other fuzzy clustering methods have been developed. In this work, we develop a fuzzy kernel k-medoids algorithm for clustering uncertain data objects. The developed fuzzy kernel k-medoids algorithm is superior to existing fuzzy k-medoids algorithms in clustering data sets with non-linearly separable clusters.

Keywords: clustering algorithm, fuzzy methods, kernel k-medoids, uncertain data

Procedia PDF Downloads 177

Authors: Tim Farrelly

Abstract:

In recent years, AI-powered object detection systems have opened the doors for innovative new applications and products, especially those operating in the real world or ‘on edge’ – namely, in sport. This paper investigates the viability of an ultra-low parameter convolutional neural network specially designed for the detection of footballs on ‘on the edge’ devices. The main contribution of this paper is the exploration of integrating new design features (depth-wise separable convolutional blocks and squeezed and excitation modules) into an ultra-low parameter network and demonstrating subsequent improvements in performance. The results show that tracking the ball from Full HD images with negligibly high accu-racy is possible in real-time.

Keywords: deep learning, object detection, machine vision applications, sport, network design

Procedia PDF Downloads 118

Authors: Yu-Shan Hsu

Abstract:

This paper studies a two-sector growth model with a technology of social constant returns and with a utility that features either a zero or a positive income effect on the demand for leisure. The purpose is to investigate how the existence of aggregate instability or equilibrium indeterminacy depends on both the intensity of the income effect on the demand for leisure and the value of the labor supply elasticity. The main finding is that when there is a factor intensity reversal between the private perspective and the social perspective, indeterminacy arises even if the utility has a positive income effect on leisure demand. Moreover, we find that a smaller value of the labor supply elasticity increases the range of the income effect on leisure demand and thus increases the possibility of equilibrium indeterminacy. JEL classification: E3; O41

Keywords: indeterminacy, non-separable preferences, income effect, labor supply elasticity

Procedia PDF Downloads 142

Authors: Muhammad Ali Sarwar, Muhammad Farooq, Nayab Hassan, Hammad Hassan

Abstract:

Pakistan is highly recognized for its agriculture and is well known for producing substantial amounts of wheat, cotton, and sugarcane. However, some factors contribute to a decline in crop quality and a reduction in overall output. One of the main factors contributing to this decline is the presence of weed and its late detection. This process of detection is manual and demands a detailed inspection to be done by the farmer itself. But by the time detection of weed, the farmer will be able to save its cost and can increase the overall production. The focus of this research is to identify and classify the four main types of weeds (Small-Flowered Cranesbill, Chick Weed, Prickly Acacia, and Black-Grass) that are prevalent in our region’s major crops. In this work, we implemented three different deep learning techniques: YOLO-v5, Inception-v3, and Deep CNN on the same Dataset, and have concluded that deep convolutions neural network performed better with an accuracy of 97.45% for such classification. In relative to the state of the art, our proposed approach yields 2% better results. We devised the architecture in an efficient way such that it can be used in real-time.

Keywords: deep convolution networks, Yolo, machine learning, agriculture

Procedia PDF Downloads 70

Authors: Ya-Mei Chang

Abstract:

This research is intended to apply spatio-temporal point process methods to the dengue fever data in Tainan. The spatio-temporal intensity function of the dataset is assumed to be separable. The kernel estimation is a widely used approach to estimate intensity functions. The intensity function is very helpful to study the relation of the spatio-temporal point process and some covariates. The covariate effects might be nonlinear. An nonparametric smoothing estimator is used to detect the nonlinearity of the covariate effects. A fitted parametric model could describe the influence of the covariates to the dengue fever. The correlation between the data points is detected by the K-function. The result of this research could provide useful information to help the government or the stakeholders making decisions.

Keywords: dengue fever, spatial point process, kernel estimation, covariate effect

Procedia PDF Downloads 324

Authors: Baha Eddine Aissou, Aichouche Belhadj Aissa

Abstract:

Light detection and ranging (LiDAR) is an active remote sensing technology used for several applications. Airborne LiDAR is becoming an important technology for the acquisition of a highly accurate dense point cloud. A classification of airborne laser scanning (ALS) point cloud is a very important task that still remains a real challenge for many scientists. Support vector machine (SVM) is one of the most used statistical learning algorithms based on kernels. SVM is a non-parametric method, and it is recommended to be used in cases where the data distribution cannot be well modeled by a standard parametric probability density function. Using a kernel, it performs a robust non-linear classification of samples. Often, the data are rarely linearly separable. SVMs are able to map the data into a higher-dimensional space to become linearly separable, which allows performing all the computations in the original space. This is one of the main reasons that SVMs are well suited for high-dimensional classification problems. Only a few training samples, called support vectors, are required. SVM has also shown its potential to cope with uncertainty in data caused by noise and fluctuation, and it is computationally efficient as compared to several other methods. Such properties are particularly suited for remote sensing classification problems and explain their recent adoption. In this poster, the SVM classification of ALS LiDAR data is proposed. Firstly, connected component analysis is applied for clustering the point cloud. Secondly, the resulting clusters are incorporated in the SVM classifier. Radial basic function (RFB) kernel is used due to the few numbers of parameters (C and γ) that needs to be chosen, which decreases the computation time. In order to optimize the classification rates, the parameters selection is explored. It consists to find the parameters (C and γ) leading to the best overall accuracy using grid search and 5-fold cross-validation. The exploited LiDAR point cloud is provided by the German Society for Photogrammetry, Remote Sensing, and Geoinformation. The ALS data used is characterized by a low density (4-6 points/m²) and is covering an urban area located in residential parts of the city Vaihingen in southern Germany. The class ground and three other classes belonging to roof superstructures are considered, i.e., a total of 4 classes. The training and test sets are selected randomly several times. The obtained results demonstrated that a parameters selection can orient the selection in a restricted interval of (C and γ) that can be further explored but does not systematically lead to the optimal rates. The SVM classifier with hyper-parameters is compared with the most used classifiers in literature for LiDAR data, random forest, AdaBoost, and decision tree. The comparison showed the superiority of the SVM classifier using parameters selection for LiDAR data compared to other classifiers.

Keywords: classification, airborne LiDAR, parameters selection, support vector machine

Procedia PDF Downloads 122

Authors: Y. Kourd, D. Lefebvre

Abstract:

The monitoring of industrial processes is required to ensure operating conditions of industrial systems through automatic detection and isolation of faults. This paper proposes a method of fault diagnosis based on a neuro-fuzzy hybrid structure. This hybrid structure combines the selection of threshold and decision tree. The validation of this method is obtained with the DAMADICS benchmark. In the first phase of the method, a model will be constructed that represents the normal state of the system to fault detection. Signatures of the faults are obtained with residuals analysis and selection of appropriate thresholds. These signatures provide groups of non-separable faults. In the second phase, we build faulty models to see the flaws in the system that cannot be isolated in the first phase. In the latest phase we construct the tree that isolates these faults.

Keywords: decision tree, residuals analysis, ANFIS, fault diagnosis

Procedia PDF Downloads 593

Authors: Xinyuan Chen, Mohd Nizam Husen, Zhongmei Zhou, Gongde Guo, Wei Gao

Abstract:

Temporal Knowledge Graph Completion (TKGC) has attracted increasing attention for its enormous value; however, existing models lack capabilities to capture both local interactions and global dependencies simultaneously with evolutionary dynamics, while the latest achievements in convolutions and Transformers haven't been employed in this area. What’s more, periodic patterns in TKGs haven’t been fully explored either. To this end, a multi-stage hybrid architecture with convolution-backed Transformers is introduced in TKGC tasks for the first time combining the Hawkes process to model evolving event sequences in a continuous-time domain. In addition, the seasonal-trend decomposition is adopted to identify periodic patterns. Experiments on six public datasets are conducted to verify model effectiveness against state-of-the-art (SOTA) methods. An extensive ablation study is carried out accordingly to evaluate architecture variants as well as the contributions of independent components in addition, paving the way for further potential exploitation. Besides complexity analysis, input sensitivity and safety challenges are also thoroughly discussed for comprehensiveness with novel methods.

Keywords: temporal knowledge graph completion, convolution, transformer, Hawkes process, periodicity

Procedia PDF Downloads 47

Authors: Anahita Mortazavi Manesh, Mojtaba Bagherzadeh

Abstract:

Metalloporphyrins are well known to mimic the activity of monooxygenase enzymes. In this regard, metalloporphyrin complexes have been largely employed as valuable biomimetic catalysts, owing to the critical roles they play in oxygen transfer processes in catalytic oxidation reactions. Investigating in this area is based on different strategies to design selective, stable and high turnover catalytic systems. Immobilization of expensive metalloporphyrin catalysts onto supports appears to be a good way to improve their stability, selectivity and the catalytic performance because of the support environment and other advantages with respect to recovery, reuse. In other words, supporting metalloporphyrins provides a physical separation of active sites, thus minimizing catalyst self-destruction and dimerization of unhindered metalloporphyrins. Furthermore, heterogeneous catalytic oxidations have become an important target since their process are used in industry, helping to minimize the problems of industrial waste treatment. Hence, the immobilization of these biomimetic catalysts is much desired. An attractive approach is the preparation of the heterogeneous catalyst involves immobilization of complexes on silica coated magnetic nano-particles. Fe3O4@SiO2 magnetic nanoparticles have been studied extensively due to their superparamagnetism property, large surface area to volume ratio and easy functionalization. Using heterogenized homogeneous catalysts is an attractive option to facile separation of catalyst, simplified product work-up and continuity of catalytic system. Homogeneous catalysts immobilized on magnetic nanoparticles (MNPs) surface occupy a unique position due to combining the advantages of both homogeneous and heterogeneous catalysts. In addition, superparamagnetic nature of MNPs enable very simple separation of the immobilized catalysts from the reaction mixture using an external magnet. In the present work, an efficient heterogeneous catalyst was prepared by immobilizing manganese porphyrin on functionalized magnetic nanoparticles through the amino propyl linkage. The prepared catalyst was characterized by elemental analysis, FT-IR spectroscopy, X-ray powder diffraction, atomic absorption spectroscopy, UV-Vis spectroscopy, and scanning electron microscopy. Application of immobilized metalloporphyrin in the oxidation of various organic substrates was explored using Gas chromatographic (GC) analyses. The results showed that the supported Mn-porphyrin catalyst (Fe3O4@SiO2-NH2@MnPor) is an efficient and reusable catalyst in oxidation reactions. Our catalytic system exhibits high catalytic activity in terms of turnover number (TON) and reaction conditions. Leaching and recycling experiments revealed that nanocatalyst can be recovered several times without loss of activity and magnetic properties. The most important advantage of this heterogenized catalytic system is the simplicity of the catalyst separation in which the catalyst can be separated from the reaction mixture by applying a magnet. Furthermore, the separation and reuse of the magnetic Fe3O4 nanoparticles were very effective and economical.

Keywords: Fe3O4 nanoparticle, immobilized metalloporphyrin, magnetically separable nanocatalyst, oxidation reactions

Procedia PDF Downloads 273