Journal of Science & Technology (JST)
Not a member yet
967 research outputs found
Sort by
Forensic Identification of the source of the morphed image: A case study on serious female disgrace.
Get PDFThe paradigm of the nature and means of functioning of crimes involving female humiliation has evolved as a result of advancements in digital image processing tools and techniques. Using digital image acquisition and transposition technology, altered images are produced that are closer to perfection. By employing various image enhancement techniques on suspected parts of the digital image and doing a pixel-by-pixel analysis of the observations obtained there by an expert in processing images, you can detect editing or morphing anomalies. In this case study, the authors provided an in-depth review of a morphing image and obtained proof that the image was taken from a website. Following website exploration, the source image was located using morphological criteria. The source image that was transformed and recognized contained the same quantization for the horizontal and vertical resolution, measured in DPI (dots per inch), and bit depth
Deep CNN Framework for Object Detection and Classification System from Real Time Videos
Get PDFIn today's world, accurately counting and classifying vehicles in real-time has become a critical task for effective traffic management, surveillance, and transportation systems. It plays a crucial role in optimizing road infrastructure, enhancing safety measures, and making informed decisions for traffic planning. With the ever-increasing traffic congestion and road safety concerns, the demand for a robust and automated vehicle counting and classification system has grown significantly. Traditionally, vehicle counting, and classification involved manual deployment of sensors or fixed cameras at specific locations. However, these methods had limitations in handling complex traffic scenarios, especially in real-time, and were less efficient in dealing with varying environmental conditions, occlusions, and different vehicle types. Fortunately, recent advancements in deep learning models have revolutionized object detection, making real-time vehicle counting and classification achievable. One such model is the YOLO (You Only Look Once) algorithm based on the Darknet framework. Leveraging the power of this model, a real-time vehicle counting, and classification system has been developed, utilizing the OpenCV library. The system employs a pretrained YOLO model to detect the number of vehicles present in a given video and classifies the type of each vehicle. By doing so, it eliminates the need for extensive human intervention and ensures automated and accurate counting of vehicles in real-time. Moreover, this system excels in handling varying traffic conditions and different vehicle types, which enhances its accuracy and reliability. The benefits of this proposed system are numerous. It provides valuable data for traffic analysis, enabling better traffic management strategies and improved infrastructure planning. With this system in place, authorities can efficiently address traffic congestion, implement targeted safety measures, and optimize traffic flow. Further, the integration of the YOLO algorithm within the Darknet framework in the proposed system has opened new possibilities for real-time traffic management. By leveraging deep learning, this system offers a reliable and efficient solution to the challenges posed by modern traffic scenarios, helping to create safer and more organized road networks for everyone
ENHANCING SEARCH ADVERTISING RECOGNITION: A COMPREHENSIVE STUDY ON FEATURE ENGINEERING TECHNIQUES AND THEIR IMPACT ON USER ENGAGEMENT
Get PDFIn the realm of digital advertising, particularly in the context of search engine advertising, businesses compete for visibility and user engagement. Search advertising recognition refers to the process of identifying relevant ads to display when a user performs a search query. The effectiveness of this recognition directly impacts the user experience and the revenue generated by advertisers and search engines. Traditional systems for search advertising recognition often relied heavily on keyword matching, bid prices, and ad quality scores. These systems used rule-based algorithms and heuristics to match user queries with relevant ads. While effective to some extent, they lacked the ability to understand the semantic context of the queries or the intent behind them. This limitation led to the development of more intelligent and adaptive systems. Thus, effective search advertising recognition is crucial for search engines like Google, Bing, or Yahoo, as well as for advertisers. Advertisers need their ads to be shown to the right audience, ensuring their investments translate into meaningful leads or sales. Users, on the other hand, rely on search engines to provide them with accurate and relevant results quickly. Therefore, this research aims to build a system with the goal is to identify the most relevant ads from a pool of available advertisements. The relevance of an ad is determined by various factors such as the semantic match between the query and the ad, historical user behavior, and the quality of the ad itself. The proposed model can accurately predict the user's intent based on the query and select ads that are not only contextually relevant but also likely to result in user engagemen
SURFACE IDENTIFICATION OF ROBOT SENSED DATA AN ARTIFICIAL INTELLIGENCE APPROACH
Get PDFIn recent years, the integration of robotics and artificial intelligence (AI) has gained significant momentum across various industries. Robots equipped with sensors play a crucial role in data acquisition for tasks such as environmental monitoring, industrial automation, and autonomous navigation. Surface identification, specifically the ability to recognize and understand the surfaces in a robot's environment, is essential for enabling precise and context-aware robotic operations. The history of surface identification in robotics is closely tied to the evolution of computer vision and machine learning. Early robotic systems relied on basic sensor data for navigation, often struggling with accurate perception of the surrounding environment. Over time, advancements in computer vision techniques and AI algorithms have enabled robots to extract meaningful information from sensor data, leading to more sophisticated capabilities, including surface identification. The challenge in surface identification for robot-sensed data lies in developing algorithms that can robustly and accurately differentiate between various surfaces in the environment. This involves recognizing and classifying different types of surfaces such as floors, walls, obstacles, and other objects. Traditional methods often face difficulties in handling complex and dynamic environments, where lighting conditions, object orientations, and material variations can affect the accuracy of surface identification. Traditional systems for surface identification in robot-sensed data often rely on rulebased approaches or simple heuristics. These methods may use thresholding techniques or predefined rules to classify surfaces based on sensor readings. However, these approaches have limitations when faced with the complexity and variability inherent in real-world environments. They may struggle with adaptability to changing conditions and lack the ability to generalize across diverse scenarios. The increasing demand for more sophisticated robotic applications underscores the need for advanced surface identification capabilities. AI approaches, particularly those leveraging deep learning and neural networks, offer the potential to significantly improve the accuracy and robustness of surface identification in robot-sensed data. An artificial intelligence approach to surface identification involves training models, such as convolutional neural networks (CNNs), on labeled datasets containing examples of different surfaces. These models can learn to automatically extract relevant features from sensor data, allowing the robot to discern and classify surfaces with greater accuracy. The use of AI in surface identification enhances adaptability, allowing robots to navigate and interact with their environment more effectively
AN ENHANCED MULTI-MODAL BIOMETRIC AUTHENTICATION SYSTEM USING MODIFIED DEEP LEARNING MODEL
Get PDFThe acceleration of the emergence of modern technological resources in recent years has given rise to a need for accurate user recognition systems to restrict access to the technologies. The biometric recognition systems are the most powerful option to date. Biometrics is the science of establishing the identity of a person through semi or fully automated techniques based on behavioural traits, such as voice or signature, and/or physical traits, such as the iris and the fingerprint. The unique nature of biometrical data gives it many advantages over traditional recognition methods, such as passwords, as it cannot be lost, stolen, or replicated. Biometric traits can be categorized into two groups: extrinsic biometric traits such as iris and fingerprint, and intrinsic biometric traits such as palm. Extrinsic traits are visible and can be affected by external factors, while the intrinsic features cannot be affected by external factors. In general, the biometric recognition system consists of four modules: sensor, feature extraction, matching, and decision-making modules. There are two types of biometric recognition systems, unimodal and multimodal. The unimodal system uses a single biometric trait to recognize the user. While unimodal systems are trustworthy and have proven superior to previously used traditional methods, but they have limitations. These include problems with noise in the sensed data, nonuniversality problems, vulnerability to spoofing attacks, intra-class, and inter-class similarity. Basically, multimodal biometric systems require more than one trait to recognize users. They have been widely applied in real-world applications due to their ability to overcome the problems encountered by unimodal biometric systems. In multimodal biometric systems, the different traits can be fused using the available information in one of the biometric system’s modules. The advantages of multimodal biometric systems over unimodal systems have made them a very attractive secure recognition method.Therefore, with the increasing demand for information security and security regulations all over the world, biometric recognition technology has been widely used in our everyday life. In this regard, multimodal biometrics technology has gained interest and became popular due to its ability to overcome several significant limitations of unimodal biometric systems. In this project, an enhanced multi-modal biometric authentication system is presented using modified deep learning model to authenticate persons using different biometric features such as Face, Iris, Finger, Palm and Ear
Green Synthesis of Zn-Ag Nanocomposites via Tamarindus indica Leaf Extracts for Antimicrobial Applications
Get PDFThe Synthesis of mental oxide nanoparticles has emerged as a burge ongoing field of research due to it’s promisingapplications in advance in technological frontiers. Particularly the utilization of biologically derived nano materials has gainedsignificant traction with in realm of nano technology Tamarindus indica plant extract has demonstrated efficacy as a viable precursorfor the synthesis of silver nano particles. Tamarindus indica being a locally abundant plant boasts a rich composition of essential aminoacids and vitamins .This investigation presents the biosynthesis of mental oxide nano particles using an aqueous extract derived fromTamarindus indic
Machine Learning Algorithms for Handwritten Devanagari Character Recognition: A Systematic Review
Get PDFDevanagari Character Recognition is a system in which handwritten Image is recognized and converted into a digital form. Devanagari handwritten character recognition system is based on Deep learning technique,
which manages the recognition of Devanagari script particularly Hindi. This recognition system mainly has five
stages i.e. Pre-processing, Segmentation, Feature Extraction, Prediction and Post processing. This paper has analyzed the approach for recognition of handwritten Devanagari characters. There are various approaches to solve this. Some of the methods along with their accuracy and techniques used are discussed here. Depending upon the dataset and accuracies of each character the techniques differs
Influence of fertigation and organic sources of nutrients on chlorophyll content and soil microbial population of Tuberose
Get PDFAn experiment was conducted to investigate the effect of fertigation, microbial consortium and biostimulants on chlorophyll content and soil microbial population in Tuberose cv. Prajwal at T.Pudhupatti village, Dindigul District of Tamil Nadu during 2015- 16 and 2016-17. The experiment was laid out in randomized block design (RBD) with nineteen treatments including each three levels of water-soluble fertilizers viz., 125, 100 and 75 per cent of the recommended doses of fertilizers along with microbial consortium, foliar spray of panchagavya and humic acid and the treatments were replicated twice. The results revealed that, 100 per cent of recommended dose of fertilizer through fertigation along with microbial consortium @ 12.5 kg ha-1, panchagavya @ 3 per cent and humic acid @ 0.4 per cent (T9) registered the highest chlorophyll “a”, chlorophyll “b”, total chlorophyll content and soil bacteria, fungi and actinomycets population. The increase of total chlorophyll by the T9 over soil application of recommended dose of fertilizer was 33 per cent
Design of Image Processing Techniques for Smart Real-Time Tracking System during Health Emergencies
Get PDFCurrently, in the period of autonomous era, every sector is involving in the adoption of systems that are spontaneous and efficient. In the sector of autonomous vehicles, safety and health of the people who are seated inside the vehicle are majorly concerned. But, the dream of complete autonomous vehicle is far. The automobile industry focuses more on surviving the accidents by making use of available tools and technologies. One such idea is discussed in this paper, which can improve the road safety by establishing better connectivity between the emergency departments and the vehicle. Under the on-road emergency condition, the proposed methodology mainly deals with, 1) Emergency message communication. 2) Automatic road side parking. In order to accomplish the above-mentioned tasks, it requires the use of sensors (Ultrasonic, Infrared and passive infrared), communication module (GPRS/GSM) operated by main (Raspberry pi) and sub-control-units (Arduino), where the main control unit supervises sub control unit and the sub control unit performs the specified tasks, as defined by the main control unit. The proposed system is developed and analyzed for a prototype car, which will be discussed further.
The paper also depicts some of the popular image processing techniques such as SVM classifier, CLAHE, histograms that can be implemented in the emergency automatic parking system to analyze the emergency situation more constructively
Semantic Dredging in the English Translation of Moved by Scholars' not Meeting with Good Fortune
Get PDF The functions of annotation in the paratext are very diverse and flexible. Semantic dredging is the most basic and common function of annotation. It undertakes the interpretation of the meaning of the original text, and tries to make cross-cultural, cross racial and cross background readers understand the original meaning in their own way of thinking. Moved by scholars' not meeting with good fortune is a work of ancient Chinese dynasties with rich literary meaning and extensive content. It is not only a challenge for English readers, but also a difficulty for Chinese readers. Therefore, it is very necessary to dredge the semantics of the paratext.