Professor (Dr.) Ratnadeep R . Deshmukh | Dr.Babasaheb Ambedkar Marathwada University (original) (raw)

Papers by Professor (Dr.) Ratnadeep R . Deshmukh

Thermal Face Recognition using Artificial Neural Network

2020 International Conference on Smart Innovations in Design, Environment, Management, Planning and Computing (ICSIDEMPC)

Face recognition is the process of identifying the human face with the help of computerized syste... more Face recognition is the process of identifying the human face with the help of computerized system mostly for the purpose of security reasons. As the face is very important characteristic of human body through which one can uniquely identify the person. So focusing the face for authentication or identification is at most important for variety of its applications. It is the problem of pattern recognition which can be easily and efficiently solved with artificial neural networks (ANN). Pattern recognition is the process through which particular patterns can be identified, recognized based on the relatedness and similarities between the patterns. Face recognition is having number of applications such as authentication at various offices, organizations or other sectors where security is of utmost important. The proposed methodology for face recognition made use of visible, thermal and fused image database. Thermal imaging detects the heat of the object or face. The artificial neural network is a very efficient technology to enhance the results. Here backpropagation and Levenberg-Marquardt algorithm were compared by using the same database. The Backpropagation algorithm achieved the accuracy of 92.86% and for Levenberg-Marquardt accuracy is 83.92%.

Spectral Biometric Verification System for Person Identification

Information and Communication Technology for Sustainable Development, 2017

Automatic person identification is possible through many biometric techniques which provide easy ... more Automatic person identification is possible through many biometric techniques which provide easy solution like identification and verification. But there may be chances of spoofing attack against biometric system. Biometric devices can also be spoofed artificially by plastic palmprint, copy medium to provide a false biometric signature, etc., so existing biometric technology can be enhanced with a spectroscopy method. In this paper, ASD FieldSpec 4 Spectroradiometer is used to overcome this problem, the palmprint spectral signatures of every person are unique in nature. Preprocessing technique including smoothing was done on the palmprint spectra to remove the noise. Statistical analysis were done on preprocessed spectra, FAR (False acceptance Rate), and FRR (False Rejection Rate) values against different threshold values were obtained and equal error rate was acquired. EER of the system is approximately 12% and the verification threshold 0.12.

The Survey of Complications in Iris Recognition

Authentication systems based on iris play important role to improve efficiency in biometric ident... more Authentication systems based on iris play important role to improve efficiency in biometric identification due to its reliability in highly secured areas. Iris recognition has been done by many researchers in last decade.The iris recognition systems have made large progress over the past decade. In this paper we will discuss about the complications and problems of iris recognition system.This paper presents an analysis of how the iris recognition is impacted by eye diseases. The challenge to improve performance of iris recognition has been issued.

Spectral Analysis of Chlorophyll, Water Content within Fresh and Water Stressed Leaves Using Hyperspectral Data

Chlorophyll is the most important content that is required for the photosynthesis process as well... more Chlorophyll is the most important content that is required for the photosynthesis process as well as one of the most important biochemical parameters of plants and is generally an indicator of plants nutritional status, photosynthetic capacity and the health status of plants; that is why, it is an important information parameter in research on crop quality monitoring, ecosystem productivity estimation, carbon cycles, etc. Water deficits can cause chlorophyll reduction which reduces the total concentration of chlorophyll. Water enables the incorporation of CO2 carbon into carbohydrates. It also produces the hydraulic pressure required for opening stomata and maintaining structural integrity. Thus, evaluating leaf water content is a key element in determining the health and productivity of vegetation. In this experiment spectral indices are used for estimating the chlorophyll and water content. The result shows high fluctuation of values between day 1and day3. With these results healt...

A Review on Marathi Speech Recognition

SSRN Electronic Journal, 2019

Indian Journal of Computer Science and Engineering, 2018

We proposed stacking ensemble model to solve the problem of feature-based opinion mining. We used... more We proposed stacking ensemble model to solve the problem of feature-based opinion mining. We used Naive Bayes, Support Vector Machine and K-Nearest Neighbor as base learner and Support Vector Machine as Meta Classifier. Using domain knowledge the dataset consists of Feature-Opinion-Negation triple is created and trained using the proposed stacking ensemble model. The proposed model predicts feature based opinion polarity identification of 4096 laptop product reviews with 92.5315% accuracy.

Improved Linear Extrapolation Technique for Crop Health Monitoring Using Hyperspectral Data

Red edge position is observed between 680 and 780 nm of spectral reflectance of green vegetation.... more Red edge position is observed between 680 and 780 nm of spectral reflectance of green vegetation. REP has a strong correlated with foliar chlorophyll content, and this is one of the key indicators of plant health. In this paper, improved linear extrapolation technique is proposed, which generalizes the processes of finding four points used for detecting red edge position. Hyperspectral reflectance of healthy and infected plants was recorded using Analytical Spectral Devices (ASD) FieldSpec Pro spectroradiometer in the range of 350–2500 nm. Reflectance and first-order derivative transform in the range 680–780 nm is used for experimentation. Results are compared with maximum first derivative, linear interpolation and linear extrapolation techniques. It is observed that improved linear extrapolation technique is more precise in detecting REP for plant health monitoring. Results are also verified by calculating hyperspectral vegetation indices of red edge region.

Removing Labeled Portion from Mammogram Image

2020 International Conference on Smart Innovations in Design, Environment, Management, Planning and Computing (ICSIDEMPC), 2020

Finding of breast tumor which is cancerous at an initial phase is essential and critical for actu... more Finding of breast tumor which is cancerous at an initial phase is essential and critical for actual treatment and patient recovery. Mammograms are specifically used only in the breast tumor identification. The images taken as mammogram ensure an ability to support physicians in detection of disease affected by cells normal growth. Each mammogram image is having label or number. The paper is based on removing an extra portion like label from mammogram images. To implement this images are used from database which is publicly available mammogram dataset provided by the Mammographic Image Analysis Society. This database contains images those are having three types of fundamental breast tissues. This implementation contains image acquisition, image de-noising, binarization and removing label.

International Journal of Recent Technology and Engineering (IJRTE), 2019

The texture of soil i.e. Sand, Silt and Clay are the most important physical properties of soil f... more The texture of soil i.e. Sand, Silt and Clay are the most important physical properties of soil for agricultural management. In the agricultural practices to increase the productivity of soil, moisture-holding capacity, aeration and to support the agronomic decisions the knowledge of soil texture is an essential task. For this purpose, the present research gives better results and fast acquisition of soil information with the use of Visible and Near Infrared (Vis- NIR) Diffuse Reflectance Spectroscopy. A total of 30 soil samples from two different locations from Aurangabad, Maharashtra, India were collected and analyzed for soil texture. To detect the soil texture the Vis-NIR DRS has shown levels of accurate results compared to the traditional laboratory method with less time, cost and effort. To measure the reflectance of soil the ASD FieldSpec4 Spectroradiometer (350-2500nm) was used. By the observation of captured spectra by using Spectroradiometer it showed that on the basis of ...

International Journal of Computer Applications, 2015

As internet usage increased users uses internet not only to access and search information but als... more As internet usage increased users uses internet not only to access and search information but also at the same time able to spread and publish own idea, sentiments, knowledge via different number of websites. Different websites encourage their user to write their views in the form of electronic text. This system increasing user-written electronic text in the world of internet, large numbers of user opinions are available on World Wide Web. User review contains important information, which is beneficial for customer as well as retailer. These reviews are in scattered format so extracting important data from this large corpus is time consuming work. Here developing a system which will automatically identify and rank the product features. The Stanford parser is used for identify product features. Sentence level sentiment classification is used for identify sentiment of each sentence separately, Sentiment Classifier is used for classifying each sentence, and finally a probabilistic ranking algorithm is used to rank the product features.

Journal of emerging technologies and innovative research, Mar 1, 2021

Rainfall is the primary key driver of an ecosystem, and it is responsible for sustaining life on ... more Rainfall is the primary key driver of an ecosystem, and it is responsible for sustaining life on the planet. Conversely, the rainfall deficiency creates a severe problem in the region, like water shortage for domestic, industrial and agricultural sectors. Furthermore, the region's land cover severely affects the barren land increment and decrement of the vegetation cover. To estimate and analyze the affected area using a manual survey is a very complex and tedious task. Therefore, meteorological and satellite imagery data provides a solution for effectively assessing the Land Use Land Cover (LULC) of the area. In the present research study, the historical rainfall data of the past 39 years and Sentinel 2 satellite data of the year 2016 and 2019 of the Vaijapur tehsil were used for the analysis. The Standard Precipitation Index (SPI) algorithm was applied to the historical rainfall dataset to investigate the region's meteorological drought. Furthermore, the ground truth points were used for the generation of training and testing pixel for classification using the Random Forest (RF) classifier. The Overall Accuracy (OA) was obtained 92% and kappa 0.9 in 2016 and 88.69% OA and 0.85 Kappa in 2019.

Abstract— Semantic Search has become a buzzword in Web Mining. Researchers have developed variety... more Abstract— Semantic Search has become a buzzword in Web Mining. Researchers have developed variety of algorithms for semantic search. Some of the methods use Search Engines hit count of a sentence for similarity measure. Example of this can be Google Distance measures. A problem of word substitution in the text can be solved by using similarity search measures. Generally, word substitution detection has gained utmost importance as terrorist groups are using substitutions for conveying their messages to their counter parts via email. As the substituted words are normal word, it is difficult to automatically recognize it. This paper discusses the methods for detection of substituted word based on search counts like Normalized Google Distance (NGD) and k-gram frequency for measurement of similarity.

Honey Adulteration Detection using Hyperspectral Imaging and Machine Learning

2022 2nd International Conference on Artificial Intelligence and Signal Processing (AISP)

Biometric Template Protection with Fuzzy Vault and Fuzzy Commitment

Proceedings of the Second International Conference on Information and Communication Technology for Competitive Strategies - ICTCS '16, 2016

Conventional security methods like password and ID card methods are now rapidly replacing by biom... more Conventional security methods like password and ID card methods are now rapidly replacing by biometrics for identification of a person. Biometrics uses physiological or behavioral characteristics of a person. Usage of biometric raises critical privacy and security concerns that, due to the noisy nature of biometrics, cannot be addressed using standard cryptographic methods. The loss of an enrollment biometric to an attacker is a security hazard because it may allow the attacker to get an unauthorized access to the system. Biometric template can be stolen and intruder can get access of biometric system using fake input. Hence, it becomes essential to design biometric system with secure template or if the biometric template in an application is compromised, the biometric signal itself is not lost forever and a new biometric template can be issued. One way is to combine the biometrics and cryptography or use transformed data instead of original biometric template. But traditional cryptography methods are not useful in biometrics because of intra-class variation. Biometric cryptosystem can apply fuzzy vault, fuzzy commitment, helper data and secure sketch, whereas, cancelable biometrics uses distorting transforms, Bio-Hashing, and Bio-Encoding techniques. In this paper, biometric cryptosystem is presented with fuzzy vault and fuzzy commitment techniques for fingerprint recognition system.

Deep Learning-Based Models for Classification of Invasive Plant Species from Hyperspectral Remotely Sensed Data

Proceedings of the International Conference on Data Science, Machine Learning and Artificial Intelligence, 2021

Invasive plant species are plants, which spread extensively outside their native ecosystem. They ... more Invasive plant species are plants, which spread extensively outside their native ecosystem. They pose a threat to the environment and economy at global and local scales, making effective mapping and detecting essential. This study aims to develop deep neural network-based classification models for hyperspectral data. The spectral reflectance of ten invasive plant leaves was collected using ASD FieldSpec4 standard Hi-Res device. Samples were collected from Dr. Babasaheb Ambedkar Marathwada university campus and Himayat Bagh garden in Aurangabad city, Maharashtra, India. Two types of deep neural networks (DNN) were applied: the first one is based on the one-dimensional convolutional neural network (1D CNN), and the second is based on the convolutional long short term memory (CNN-LSTM). We proposed and compare the performance of the two models with three existing models, multilayer perceptron (MLP), random forest (RF), and support vector machine (SVM). The CNN-LSTM model achieves the highest discrimination accuracy among all the other models with an overall test accuracy of 99.3% and an F1_score of 0.98. Moreover, the CNN model achieves a high classification accuracy of 97.3% and an F1 score of 0.97, which contains the inception layer. In the CNN-LSTM model, the convolutional layer extracts the input vector of spectral features and feeds them to the LSTM layer to capture contextual information. In 1D CNN, the inception layer concatenates the output of multiple kernel sizes and adds flexibility to the model to improve the classification performance. This study revealed that our proposed models provide better performance than traditional machine learning methods and simple deep learning ones.

A Review on Automatic Classification of Honey Botanical Origins using Machine Learning

2021 International Conference of Modern Trends in Information and Communication Technology Industry (MTICTI), 2021

Honey botanical origin classification is essential to honey authentication and honey botanical or... more Honey botanical origin classification is essential to honey authentication and honey botanical origin mislabeling prevention. Recently, several researchers have used advanced analytical techniques for classifying honey floral sources. These methods have incorporated different acquisition technologies and machine learning (ML) models. In this paper, we review state-of-the-art approaches for classifying honey botanical sources. We discuss the various technologies used for measuring honey constituents, honey physical and chemical properties, and technologies for capturing honey spatial and spectral data. Also, we discuss the ML techniques and their classification performances. We give recommendations for future work at the end of this paper.

Statistical Analysis of WLR(Wheat Leaf Rust) Disease using ASD FieldSpec4 Spectroradiometer

2018 3rd IEEE International Conference on Recent Trends in Electronics, Information & Communication Technology (RTEICT), 2018

Remote sensing(RS) is the process of obtaining an information about an object(living/non-living) ... more Remote sensing(RS) is the process of obtaining an information about an object(living/non-living) which is located at distant location by sending and receiving electromagnetic waves that the object reflects or transmits back using some sensors or some information capturing devices. The applications of remote sensing techniques in the field of agriculture are wide and varied ranging from crop identification, detection of disease on different crops & predicting grain yield of crops. It typically involves measuring reflectance of electromagnetic radiation in the visible, near-infrared (NIR), or middle-infrared ranges using spectrometers. This research paper Focuses on study of Winter Wheat crop and its associated Wheat Leaf Rust (WLR) Disease by hyperspectral measurements and statistics using ASD Fieldspec4 Spectroradiometer. The range of wavelength taken for which WLR Disease infection has observed is 450 nm to 1000 nm.

Face verification using scale invariant feature transform with template security

2017 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), 2017

Biometrie is a reliable and convenient way for person authentication. Template security is a majo... more Biometrie is a reliable and convenient way for person authentication. Template security is a major concern in practical implementation of biometric system. This work proposes cryptosystem for face verification with fuzzy vault. Here, a secrete value k is locked using fuzzy vault with a set of points. Fuzzy vault scheme is order invariant and applied for point set, hence SIFT is suitable feature extraction method. SIFT is distinctive invariant feature extraction method. As there is a trade-off between accuracy and security, this is a challenge to implement biometric template protection scheme which will give better accuracy of the biometric system.

Analyzing the Effect of Database Dimensionality on Performance of Adaptive Apriori Algorithm

Communications in Computer and Information Science, 2018

Obtaining frequent itemsets from the dataset is one of the most promising areas of data mining. T... more Obtaining frequent itemsets from the dataset is one of the most promising areas of data mining. The Apriori algorithm is one of the most important algorithms for obtaining frequent itemsets from the dataset. But the algorithm fails in terms of time required as well as number of database scans. Hence a new improved version of Apriori is proposed in this paper which is efficient in terms of time required as well as number of database scans than the Apriori algorithm. It is well known that the size of the database for defining candidates has great effect on running time and memory need. The usefulness of the adaptive apriori algorithm in terms of dimensionality of the dataset is demonstrated. We presented experimental results, showing that the proposed algorithm always outperform Apriori. To evaluate the performance of the proposed algorithm, we have tested it on Turkey student’s database of faculty evaluations.

“Biometrics — Iris recognition system” A study of promising approaches for secured authentication

2016 3rd International Conference on Computing for Sustainable Global Development (INDIACom), 2016

To automatically identify and authenticate an individual person biometric systems use unique feat... more To automatically identify and authenticate an individual person biometric systems use unique features or characteristics of an individual. Since the last decade researchers are using Iris recognition for this purpose. Iris recognition technology is in more demand due to reliability and perfection towards recognition rates. Communication networks and mobile commerce are the application areas where biometrics is used to authenticate a person. To authenticate users and transactions, Iris based applications use infrared and video cameras. For identification on large scale database applications accuracy, speed and template size are important attributes. This paper outlines research made on characteristics of Iris Recognition Technology which makes it very attractive for using as an authenticating system to identify individual. The segmentation process used to locate the iris structure affects the performance of iris recognition system. The techniques which are employed for enhancing the ...