Computer Science And Technology - Research Publications
Permanent URI for this collectionhttps://kr.cup.edu.in/handle/32116/82
Browse
19 results
Search Results
Item Classification of Breast Cancer Mammographic Images Using A Light-Weighted Convolutional Neural Network(Institute of Electrical and Electronics Engineers Inc., 2023-04-10T00:00:00) Kaur, Palwinder; Kaur, AmandeepDeep learning is a method demanded by radiologists to assist them interpret and classify medical images correctly. A Convolutional Neural Network (CNN) is the most widely used method for classifying and analysing images. In this paper, a light-weighted CNN is presented for breast cancer classification using a dataset of breast mammography images. The suggested methodology improves the classification of mammary cancer images to assist radiologists in the detection of mammary cancer. The application of the proposed model can help in the diagnosis of mammary cancer using digital mammograms without any preceding information about the existence of a cancerous lesion. The proposed CNN can categorize the input medical images as malignant or benign with an accuracy of 99.35% which is the highest accuracy achieved for such a large mammography dataset. � 2023 IEEE.Item Review of Progress in Diagnostic Studies of Autism Spectrum Disorder Using Neuroimaging(Springer Science and Business Media Deutschland GmbH, 2023-01-12T00:00:00) Kaur, Palwinder; Kaur, AmandeepThis review article summarizes the recent advances in the diagnostic studies of autism spectrum disorders (ASDs) considering some of the most influential research articles from the last two decades. ASD is a heterogeneous neurodevelopmental disorder characterized by abnormalities in social interaction, communication, and behavioral patterns as well as some unique strengths and differences. The current diagnosis systems are based on autism diagnostic observation schedule (ADOS) or autism diagnostic interview-revised (ADI-R), but biological markers are also important for an effective diagnosis of ASDs. The amalgamation of neuroimaging techniques, such as structural and functional magnetic resonance imaging (sMRI and fMRI), with machine-learning and deep-learning approaches helps throw new light on typical biological markers of ASDs at the early stage of life. To assess the performance of a deep neural network, we develop a light-weighted CNN model for ASD classification. The overall accuracy, precision, and F1-score of the proposed model are 99.92%, 99.93% and 99.92%, respectively. All the neuroimaging studies we have reviewed can be divided into 3 categories, viz. thickness, volume and functional connectivity-based studies. We conclude with a discussion of the major findings of considered studies and promising directions for future research in this field. Graphical Abstract: [Figure not available: see fulltext.] � 2023, International Association of Scientists in the Interdisciplinary Areas.Item Identification of Counterfeit Indian Currency Note using Image Processing and Machine Learning Classifiers(Institute of Electrical and Electronics Engineers Inc., 2023-03-27T00:00:00) Sharan, Vivek; Kaur, Amandeep; Singh, ParvinderTechnology is continuously changing our life. Day by day, it makes our life easy, but some challenges and issues exist. Counterfeit currency is one of them. It happens because of the production and circulation of currency without the permission of an authorized system. Some people use scanning and printing technology to produce such notes and circulate them around us, which is a kind of forgery.It leads to personal loss and degrades the Country's economy. Such notes are very similar to the original, which becomes a problem for ordinary people to identify the authenticity of the currency, especially for visually impaired people. However, most researchers proposed different methods to differentiate real notes from fake ones based on the currency's shape, colors, and size. The notes are easily detected when the condition of the note is good but difficult when it deteriorates over time. Extracting features from such notes is another challenging task. Some systems are available to only specific sectors and not easily available to common people. Therefore, a system that can distinguish between real and fake notes is required. This article classifies Indian Currency notes as real or fake with four supervised Machine Learning algorithms i.e., Support Vector Classifier, K-Nearest Neighbor, Decision Tree, and Logistic Regression followed by Image Processing techniques. This study has implemented these algorithms on the dataset of 1372 currency image samples, out of which 762 are real, and the remaining are fake, available on the UCI machine learning repository. Further, the performance of all the algorithms is measured in terms of accuracy, recall, Precision, and F1-score and after analyzing, it is observed that the K-Nearest Neighbor leverage outstanding results compared to other algorithms. � 2023 IEEE.Item A systematic review of artificial intelligence in agriculture(Elsevier, 2022-01-15T00:00:00) Singh, Parvinder; Kaur, AmandeepThe current world population is 7.8 billion and is projected to reach 9.8 billion by 2050. The limited land area and strong need to produce more crop to feed the ever-increasing population is a major challenge today, especially for developing countries. The strong need to produce more crop from lesser land has led to several challenges in the field of agriculture. Reduction in agriculture yield due to climate change and global warming due to farming has become a vicious circle. Excessive use of chemicals in farms to increase soil fertility and reduce weeds and pests have adversely affected the environment and the human health. There is limited availability of natural resources like phosphorous and energy required in agriculture. Water scarcity and increase in plant diseases are other major concerns. Artificial intelligence (AI) has emerged as a promising technology in digital agriculture. Digital agriculture relates to using digital technologies for collecting, storing, and further analyzing the electronic agricultural data for better reasoning and decision-making using AI techniques. Precision agriculture is one such technique that monitors soil moisture and composition, temperature, and humidity and determines optimized fertilizer and water requirements for a specific crop and different areas of a farm. Then there are computer vision and machine learning techniques to detect diseases and deficiencies in plants, recognizing weeds that helps in spraying only those parts of land where the plants are disease-infected or where weeds are present instead of the whole field. Utilization of AI in agriculture is helping in developing agricultural methods capable of increasing crop yield and reducing the previously stated challenges. With merits of using AI, there are certain issues. The first major issue in using the AI techniques is the need for high computational power that, again, leads to global warming. Also, in developing countries, the internet infrastructure needs to be improved to use AI techniques effectively. Cost of using AI is high, and countries need AI experts to use the techniques to full potential. The focus of this chapter is to review how AI techniques are helping in increasing yield and overcoming limitations, like global warming, excessive use of fertilizers, limited availability of natural resources, plant disease, and water scarcity. The chapter concludes by discussing the issues and challenges in using AI, especially as it related to agriculture. � 2022 Elsevier Inc. All rights reserved.Item Face detection in still images under occlusion and non-uniform illumination(Springer, 2021-01-26T00:00:00) Kumar, Ashu; Kumar, Munish; Kaur, AmandeepFace detection is important part of face recognition system. In face recognition, face detection is taken not so seriously. Face detection is taken for granted; primarily focus is on face recognition. Also, many challenges associated with face detection, increases the value of TN (True Negative). A lot of work has been done in field of face recognition. But in field of face detection, especially with problems of face occlusion and non-uniform illumination, not so much work has been done. It directly affects the efficiency of applications linked with face detection, example face recognition, surveillance, etc. So, these reasons motivate us to do research in field of face detection, especially with problems of face occlusion and non-uniform illumination. The main objective of this article is to detect face in still image. Experimental work has been conducted on images having problem of face occlusion and non-uniform illumination. Experimental images have been taken from public dataset AR face dataset and Color FERET dataset. One manual dataset has also been created for experimental purpose. The images in this manual dataset have been taken from the internet. This involves making the machine intelligent enough to acquire the human perception and knowledge to detect, localize and recognize the face in an arbitrary image with the same ease as humans do it. This article proposes an efficient technique for face detection from still images under occlusion and non-uniform illumination. The authors have presented a face detection technique using a combination of YCbCr, HSV and L � a � b color model. The proposed technique improved results in terms of Accuracy, Detection Rate, False Detection Rate and Precision. This technique can be useful in the surveillance and security related applications. � 2021, The Author(s), under exclusive licence to Springer Science+Business Media, LLC part of Springer Nature.Item A new binarization method for degraded document images(Springer Science and Business Media B.V., 2019-09-17T00:00:00) Rani, Usha; Kaur, Amandeep; Josan, GurpreetThe binarization of image is an important stage in any document analysis system such as OCR. It converts the colored or grayscale images into monochromatic form to reduce the computational complexity in the next stages. In old document images in the presence of degradations (ink bleed, stains, smear, non-uniform illumination, low contrast, etc.) the separation of foreground and background becomes a challenging task. Most of the existing binarization techniques can handle only a subset of these degradations. We present a simple binarization method for old document images. The experimental results confirm that the proposed technique gives good binarization results in the presence of various degradations. It computes the Laplacian of an image to separate the foreground. The subtracted Laplacian image is binarized using a global threshold. Finally, the postprocessing using morphological functions is applied. The results are compared in terms of F-measure, PSNR, time complexity, and OCR based evaluations which shows that our method outperforms existing techniques like Niblack, Sauvola, Gatos, Zhou, NICK, Singh, and Bataineh. � 2019, Bharati Vidyapeeth's Institute of Computer Applications and Management.Item Handover between Wi-Fi and WiMAX technologies using GRE tunnel(Springer, 2019) Aimen, Aroof; Hamid,, S; Ahmad, S; Chisti, M.A; Khurana, Surinder Singh; Kaur, AmandeepThe next era of wireless grid inclines to be heterogeneous in the composition, i.e., wireless technologies like Wi-Fi and WiMAX networks desire co-breathe, so there is a demand for the best utilization of the accessible mixed chains. This paper considers the issue of handover between Wi-Fi and WiMAX grids with seamless connectivity. For this, first, a mobile terminal that abuts both IEEE 802.11 and IEEE 802.16 technologies was designed in the simulator. The developed mobile node was then introduced in the simulation scenario to study the various metrics. Second, we present the incorporation of GRE tunnel between the home agent and base stations for doing away with latency and packet drop and thereby improving the MOS value of the interest of consumers, giving impetus to efficiency day instant and day forth. © 2019, Springer Nature Singapore Pte Ltd.Item Face detection techniques: a review(Springer, 2019) Kumar, A; Kaur, Amandeep; Kumar, M.With the marvelous increase in video and image database there is an incredible need of automatic understanding and examination of information by the intelligent systems as manually it is getting to be plainly distant. Face plays a major role in social intercourse for conveying identity and feelings of a person. Human beings have not tremendous ability to identify different faces than machines. So, automatic face detection system plays an important role in face recognition, facial expression recognition, head-pose estimation, human–computer interaction etc. Face detection is a computer technology that determines the location and size of a human face in a digital image. Face detection has been a standout amongst topics in the computer vision literature. This paper presents a comprehensive survey of various techniques explored for face detection in digital images. Different challenges and applications of face detection are also presented in this paper. At the end, different standard databases for face detection are also given with their features. Furthermore, we organize special discussions on the practical aspects towards the development of a robust face detection system and conclude this paper with several promising directions for future research. © 2018, Springer Nature B.V.Item A Comprehensive Review on Performance of AODV and DSDV Protocol Using Manhattan Grid Mobility Model(IJRET, 2014) Kaur, Amandeep; Mittal, MeenakshiWireless networks have become an epitome of revolution in the communication industry as these have enabled the devices to communicate and access information independent of their location. These networks can be classified into two categories: Infrastructure based and Infrastructure less. Mobile ad hoc networks (MANET) fall under infrastructureless category in which nodes are able to move thereby making the topology of the network highly dynamic. Due to the dynamically changing topology, efficient routing mechanisms needed to be developed, which led to the foundations of various mobile ad hoc routing protocols. There are a number of mobile ad hoc routing protocols proposed to serve different purposes like security and transmission efficiency. These protocols are divided into two categories: Table based and Demand based. Through this work, table based traditional routing protocol DSDV and demand based routing protocol AODV have been assessed through simulation using Manhattan Grid mobility model. Comprehensive analysis was carried out to analyze which protocol performs better in the assumed scenarios. The performance metrics evaluated for the two protocols are Throughput, Average End to End delay, Routing Overhead and Packet Delivery RatioItem Detection of defective printed circuit boards using image processing(Inderscience Publishers, 2018) Kaur, Beant; Kaur, Gurmeet; Kaur, AmandeepManufacturing of printed circuit boards involves three stages (printing, component fabrication over surface of printed circuit boards, soldering of components), where inspection at every stage is very important to improve the quality of production. Image subtraction method is widely used for finding the difference between any two images. Using this method, defects have been detected by finding the difference between reference (defect free) and test image (to be inspected). The major limitation of image subtraction is that both the images should have same size and same orientation. The proposed method removes the above explained limitation of image subtraction method and also calculates total number of defects on printed circuit boards. The proposed method is tested on six test images. The experimental results show that proposed method is simple, economical and easy to implement in small and medium scale industries where most of the inspection is still done by humans.