A5069519791- Handwritten Text Recognition Techniques
- Evolutionary Algorithms and Applications
- Face and Expression Recognition
- Metaheuristic Optimization Algorithms Research
- Text and Document Classification Technologies
- Advanced Image and Video Retrieval Techniques
- Gene expression and cancer classification
- Image Retrieval and Classification Techniques
- Advanced Multi-Objective Optimization Algorithms
- Machine Learning in Bioinformatics
- Vehicle License Plate Recognition
- Machine Learning and Data Classification
- Emotion and Mood Recognition
- Neural Networks and Applications
- Fuzzy Logic and Control Systems
- Image Processing and 3D Reconstruction
- Currency Recognition and Detection
- Electricity Theft Detection Techniques
- Marine animal studies overview
- Blind Source Separation Techniques
- Artificial Immune Systems Applications
- Underwater Acoustics Research
- Advanced Computing and Algorithms
- Underwater Vehicles and Communication Systems
- Music and Audio Processing
Jadavpur University
2017-2021
With the rapid growth in credit card based financial transactions, it has become important to identify fraudulent ones. In this work, a two stage model is proposed such transactions. To make fraud detection system trustworthy, both miss and false alarms are minimized. Understanding learning complex associations among transaction attributes major problem. address issue, at first of an autoencoder used transform feature vector lower dimension. The thus obtained as input classifier second...
Nowadays, researchers aim to enhance man-to-machine interactions by making advancements in several domains. Facial emotion recognition (FER) is one such domain which have made significant progresses. Features for FER can be extracted using popular methods. However, there may some redundant/irrelevant features feature sets. In order remove those that do not any impact on classification process, we propose a selection (FS) technique called the supervised filter harmony search algorithm (SFHSA)...
Abstract Feature selection (FS) is a technique which helps to find the most optimal feature subset develop an efficient pattern recognition model under consideration. The use of genetic algorithm (GA) and particle swarm optimization (PSO) in field FS profound. In this paper, we propose insightful way perform by amassing information from candidate solutions produced GA PSO. Our aim combine exploitation ability with exploration capacity We name new as binary (BGSO). proposed method initially...
Abstract In India, which has numerous officially recognized scripts, there is a primary need for categorizing the documents on basis of scripts used therein. Identification script in document essential its effective handling both manually and digitally. image an important research problem pattern recognition field, which, at times, suffers from issue growing dimensionality feature vector requires efficient selection technique. Keeping this fact mind, paper, we propose clustering‐based filter...
Abstract The feature selection process is very important in the field of pattern recognition, which selects informative features so as to reduce curse dimensionality, thus improving overall classification accuracy. In this paper, a new approach named Memory-Based Histogram-Oriented Multi-objective Genetic Algorithm (M-HMOGA) introduced identify subset be used for problem. proposed M-HMOGA applied two recently sets, namely Mojette transform and Regional Weighted Run Length features....
Abstract In any multi-script environment, handwritten script classification is an unavoidable pre-requisite before the document images are fed to their respective Optical Character Recognition (OCR) engines. Over years, this complex pattern problem has been solved by researchers proposing various feature vectors mostly having large dimensions, thereby increasing computation complexity of whole model. Feature Selection (FS) can serve as intermediate step reduce size restricting them only...
Abstract Word searching or keyword spotting is an important research problem in the domain of document image processing. The solution to said for handwritten documents more challenging than printed ones. In this work, a two-stage word schema introduced. first stage, all irrelevant words with respect search are filtered out from page image. This carried using zonal feature vector, called pre-selection along rule-based binary classification method. next step, holistic recognition paradigm used...