Sajid PM, Kshithy Ravindran, C Dharsana et al. · Journal of Information Technology and Digital World · 2023
Exams are the methods adopted by educational institutions to identify student’s knowledge. Students adopt various ways to cheat in exams like answer sheet exchanging, copying etc, students cheat their way into getting good grades. Detection of cheating manually may not be efficient to identify and prevent cheating during examinations. So, to avoid this the process of invigilation is made automatic. Automated invigilation offers the best method for keeping an eye on the kids and spotting instances of malpractice right away. The proposed work has three phases. In the first phase, the exam management does processes like publishing time table, allocating exam hall, allocating hall to staff etc. In the second phase the posture detection of the student present in the exam hall is done using Computer Vision and Media Pipe to detect whether the student has involved in the malpractice. In the third phase, the emotion analysis and face recognition of the student is done using the Haar Cascade Frontal Algorithm. The proposed work also helps to eliminate impersonation in the exam hall.
Cited by 0Open access
QIAO Jie, CAI Ruichu, HAO Zhifeng · SHILAP Revista de lepidopterología · 2022
The existing cascade nonlinear Additive Noise Model(ANM) can infer the causal direction of hidden intermediate variables, but fail to deal with global structure search and equivalence class recognition in the case of causal network learning that includes hidden variables and cascade causality transferring.This paper presents a two-stage causal structure learning algorithm for non-chronological observation data.In the first stage, a basic causal network skeleton is constructed based on the conditional independence between the observation data variables. In the second stage, by using a cascaded nonlinear ANM, the causal direction of the edge likelihood under the assumptions of different causal directions is inferred by comparing each adjacent causality in the skeleton.The experimental results show that the algorithm has outstanding performance on the virtual causal structure dataset for a varying number of hidden variables, average in-degree, structural dimension, and number of samples.Furthermore, the F1 value of this algorithm on the real causal structure dataset improved by 51% on average compared with mainstream causal structure learning algorithms, displaying a higher accuracy and robustness.
Cited by 0Open access
Farzaneh Ezazi, Mohammad Hassan Mallah, Javad Karimi‐Sabet et al. · Progress in Nuclear Energy · 2019
Cited by 22
Ruth Ramya Kalangi, Suman Maloji, P. Syam Sundar et al. · 2022 4th International Conference on Smart Systems and Inventive Technology (ICSSIT) · 2022
The most difficult issue in the regions of picture preparation is human face identification [1]. Rough Haar utilized for human face area alongside an additional three weak classifiers is proposed in this paper. Classifiers distinguish human faces utilizing skin conceal histogram organizing, eyes area and mouth ID [2]. The crude-Haar classifier is used to classify images of individuals [3]. A frail classifier is used to extract the dominant art of non-human faces based on face skin shade histogram. Eyes in the human face are identified using the powerless classifier. Finally, the mouth is also identified using the same classifier. OpenCV is used to produce results on images of individuals. Experimental results show that the proposed method is successful and extracts many features from the human face, and is effective.
Cited by 6
Rentao Gu, Xiaoxu Liu, Hui Li et al. · China Communications · 2013
In this paper, we propose a mathematical model for long reach Passive Optical Networks (PON) planning. The model considers the traffic demand, user requirements and physical constraints. It can support conventional star-like topologies as well as cascade PON networks. Then a two-stage evolutional algorithm is described to solve this problem. The first stage was to find a proper splitter candidate site set, composing the outer loop. The second stage aimed to get the optimal topology when the splitter locations were selected, composing the internal loop. In this algorithm, the Prüfer sequence is used to build up a one-to-one correspondence between a PON network configuration and a chromosome. Compared with the results obtained by the enumeration method, the proposed model and algorithm are shown to be effective and accurate.
Cited by 6
Amrita Tuteja, Amita Mahor · 2013
Cited by 0
G. Zakaria, A.A. Beex · 2003
We analyze the convergence behavior of the CRLS-SA algorithm for inverse filtering. The CRLS-SA is a cascade adaptive filter based on the RLS algorithm, with each section adapted independently based on global minimization. The subsection adaptation results in reduced computational complexity. The rate of convergence is evaluated based on the convergence time constant defined as the ratio of condition number and sensitivity. The smaller the convergence time constant, the faster the structure converges. Analysis and simulation explain and show that CRLS-SA exhibits faster convergence than the direct form RLS adaptive filter for speech type signals.
Cited by 1
Sabbir Ahmad, A. Antoniou · 2006
A genetic-algorithm (GA) based optimization approach for the design of multiplierless FIR filters is presented. The approach exploits a recently-introduced GA, called orthogonal GA (OGA), based on the so-called experimental design technique to obtain fixed-point implementations of linear-phase FIR filters. In the proposed approach, the filter coefficients are treated as chromosomes to be optimized and the effects of finite word length are minimized by considering the filter as a cascade of two sections. The OGA employs an integer encoding scheme for chromosome construction whereby the filter coefficients are expressed in terms of sums of powers of two to obtain the multiplierless design. The optimization is carried out by minimizing an objective function based on the amplitude response error. Experimental results show that the OGA approach leads to improved amplitude response relative to that of an equivalent direct-form cascade filter obtained using the Remez exchange algorithm
Cited by 43
Yuliadi Erdani, Afaf Fadhil Rifa’i, Agil Calfarera · 2021
Walking speed is widely used as the material for research studies in the health sector, especially for early detection of someone who is healthy or sick. The existence of a monitoring system will greatly help a company to detect the health of its workers early. In this paper, an algorithm for calculating a person’s walking speed is developed. This study uses HAAR Cascade Classifier as pedestrian detection and the process of image processing is performed with the help of the OpenCV library. The test results show that the system successfully detects humans both indoors and outdoors in dim lighting conditions (in the afternoon) and bright (during the day). Based on the test results, the system has successfully detected both single-object and multi-object human objects with accuracy in calculating running speed in the range of 86%.
Cited by 3
В. А. Палкин · Atomic Energy · 2024
Cited by 0
Keyang Luo, Yiwei Fan · 2024
Support Vector Machines (SVMs) are widely used machine learning algorithms, recognized for their strong performance across various classification tasks. However, their computational complexity, particularly with large datasets, poses significant challenges. Parallel learning offer partial relief, with the Cascade SVM being a prominent approach that partitions data into smaller subsets to manage computational load. Despite its advantages, the current Cascade SVM algorithm suffers from accuracy loss and inefficient convergence. In this paper, we present an improved parallel optimization algorithm for the Cascade SVM. Specifically, we design a novel exchange system between support vectors (SVs) and non-SVs, combining with screening rules to ensure the preservation of all global SVs, which significantly improves classification accuracy and accelerates convergence. Experimental results on synthetic and real-world datasets demonstrate that the proposed algorithm achieves accuracy comparable to that of a standard SVM, while offering substantial speed improvements over the traditional multi-pass Cascade SVM, particularly in large-scale and high-dimensional settings. Our method offers a robust and efficient solution for large-scale SVM training.
Cited by 0
Qinchuan Wang, Bin Li, Ziqian Deng et al. · 2023
In order to improve the crack detection accuracy of concrete dam surface, a cascade crack detection algorithm based on deep learning and prior information reinforcement is proposed. In the algorithm, two neural networks are cascaded to detect cracks by a cascade detection method. Prior information strengthening includes standardization and edge strengthening. When making standardized crack slices, the crack is rotated to the horizontal trend and the center of mass is placed in the section center. Meanwhile, slices belonging to the general background are removed according to the gray gradient similarity. The slices were processed with the prior information of crack edges to artificially highlight the features and reduce the difficulty of network training. The network results without prior information reinforcement and cascade detection are compared respectively to evaluate the network detection performance. The experimental results show that the accuracy rate and recall rate of concrete dam surface crack detection have been improved significantly, which proves the effectiveness of the algorithm.
Cited by 1
Yanjiao Gao, Xuyang Wang, Qinghua Li · 2023
Cervical cell early detection can timely discover abnormal cells and take treatment to avoid cervical cancer. With the development of deep learning technology, cervical cell smear detection technology has made rapid progress. However, current deep learning segmentation algorithms still have problems with inaccurate segmentation and low segmentation accuracy when dealing with high overlap of cervical cells. Therefore, we propose the cell overlap degree to measure the degree of cell overlap, in order to compare the segmentation algorithm more targetedly. At the same time, to address the misclassification problem of the traditional Cascade Mask RCNN algorithm, we propose an improved algorithm based on Cascade Mask RCNN. This algorithm uses a more advanced backbone network to improve the segmentation accuracy of the algorithm, introduces a cross-scale feature fusion structure to improve the network’s ability to extract cell cytoplasm edge feature information, and improves the ROI pooling method to further enhance the algorithm’s feature extraction ability. The improved algorithm is compared in experiments on the dataset, and the results show that our algorithm has certain effectiveness and progressiveness.
Cited by 3
Clarance Prajwan Dsouza, Likhitha D.K., B S Nausheeda · 2020
Facial recognition is the biometric technology used in face detection. With the expansion of modern society, the demand for facial identification has been honestly vital. Face detection and recognition technology have advanced globally. There are various facial detection methods. It is due to the demand for safety, which includes permission, national protection, and other crucial circumstances. This paper will outline various face reputation strategies. Haar Cascade and local Binary sample caressed for the classification. As a result, Haar Cascade's accuracy is higher than that of the nearby Binary sample, but its execution time is longer than that of the nearby Binary pattern.
Cited by 0Open access
S Kishore, Mathi Shankar S, J Joncy · International Journal for Research in Applied Science and Engineering Technology · 2023
Abstract: This paper focuses on the development of Wifi Enabled in Campus Surveillance System Using Haar Cascade Algorithm. The system captures and stores images of individuals accessing a building and matches them against a pre-existing database of authorized personnel. If a match is found, the system logs the entry time. Similarly, during exit, the same process will be followed. Unlike access cards or passwords, which can be easily stolen or shared, a person's face cannot be replicated or forged, making it a more reliable means of identification and identify any security breaches. With this system, authorized personnel can simply walk into the building, and their entry is automatically recorded. Overall, This system utilizing face recognition technology is a powerful security solution that offers enhanced security, improved efficiency, and real-time monitoring capabilities. Its ability to authenticate individuals based on their facial features is making it an attractive option for organizationslooking to improve their security.
Cited by 0Open access
Gou‐Jen Wang, Jai-Juin Chen · 2003
In the article, a new and efficient multilayer neural networks learning algorithm is presented. The key concept of this new algorithm is the two-stage implementation of the steepest descent method. At the first stage, it is used to search the optimal learning constant /spl eta/ and momentum term /spl alpha/ for each weights updating process. At the second stage, the Delta learning rule is then employed to modify the connecting weights in terms of the optimal /spl eta/ and /spl alpha/. Computer simulations show that the proposed new algorithm outmatches other learning algorithms both in convergence speed and success rate. On real industrial application, a self-tuning neural-network based PID controller for precise temperature control of an injection mode barrel system by using the developed algorithm is developed. Experiments show that the proposed self-tuning PID controller can precisely control the barrel temperature within /spl plusmn/0.5/spl deg/C.
Cited by 0
Bodyanskiy, Yevgeniy, Viktorov, Yevgen · Bulgarian Digital Mathematics Library (BulDML) at IMI-BAS (Institute of Mathematics and Informatics) · 2010
In the paper learning algorithm for adjusting weight coefficients of the Cascade Neo-Fuzzy Neural \nNetwork (CNFNN) in sequential mode is introduced. Concerned architecture has the similar structure with the \nCascade-Correlation Learning Architecture proposed by S.E. Fahlman and C. Lebiere, but differs from it in type of \nartificial neurons. CNFNN consists of neo-fuzzy neurons, which can be adjusted using high-speed linear learning \nprocedures. Proposed CNFNN is characterized by high learning rate, low size of learning sample and its \noperations can be described by fuzzy linguistic “if-then” rules providing “transparency” of received results, as \ncompared with conventional neural networks. Using of online learning algorithm allows to process input data \nsequentially in real time mode.
Cited by 4Open access
Zhongwei Li · 2006
To apply Support Vector Machine (SVM) to deal with larger training data, a training algorithm based on cascade structure is proposed, which is not based on solving a complex quadratic optimization problem but divide and conquer strategy. Cascade structure is applied to reduce the number of training data in each training process, and multiple SVM classifiers are obtained which represented learning results of every training subset. The support vector sets obtained correspondingly are combined and added back into training subsets as feedbacks. Feedbacks are necessary when considering the problem that the learning results are subject to the distribution state of the training data in different subsets. The experimental results on UCI dataset show that the proposed training algorithm is able to deal with larger scale learning problems, and the suitable feedback strategy makes the learning accuracy more satisfying and less computation time cost compared with standard cascade SVM algorithm.
Cited by 4
N. Dileep kumar, S. Shanthi · International Journal of Engineering and Advanced Technology · 2020
Now a day, in every single person households it is important to check regularly regarding their safety. Especially for elderly people it is mandatory, because they have become a target for certain burglars which leads to higher accidents/robberies in almost all the areas. To decrease the risk of such unwanted happenings in living space for single-person households, the hybrid security system should be adopted. The automatic personal identification has become the popular instead of using passwords or pattern in this days. This paper addresses the development of a face recognition technique for the above mentioned purpose.
Cited by 4Open access
Satyajit S. Uparkar, Ujwal A. Lanjewar · 2022 IEEE World Conference on Applied Intelligence and Computing (AIC) · 2022
As a result of the huge amounts of data available in any subject of research, data mining technologies confront considerable obstacles. Due to the large volume of data, most existing data mining methods are inapplicable to many real-world problems. Data mining methods become ineffective when the problem size becomes too large. Scalability is an issue that must be addressed in data mining algorithms in order to construct high-performance, efficient, and scalable data mining algorithms. A new scalability technique is proposed and applied to several data mining problems in this research. The proposed scaling methodology is developed using a cascade approach. The approach begins with the collection of a large data set from several sources, which is then preprocessed. Once the dataset has been preprocessed, decompose it into smaller data sets of equal size subsets. Then apply a data mining approach to each subset, with the identical data mining method stated for each subset. The findings of the data mining approach on all subsets are pooled and aggregated for the final output. The performance of the suggested algorithm is assessed using a number of criteria, including accuracy, precision, recall, F-score, and execution time. Social advertising and bank marketing are the two datasets on which the proposed method is tested. The suggested approach's performance is compared with non-scale data mining methods, and it is found that the scaling method outperformed non-scale data mining methods on all measures.
Cited by 9