Bibhas Chandra Dhara

ORCID: 0000-0003-3731-0005
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Advanced Steganography and Watermarking Techniques
  • Chaos-based Image/Signal Encryption
  • Biometric Identification and Security
  • Digital Media Forensic Detection
  • Speech and Audio Processing
  • Music and Audio Processing
  • Advanced Data Compression Techniques
  • User Authentication and Security Systems
  • Retinal Imaging and Analysis
  • Music Technology and Sound Studies
  • Face recognition and analysis
  • Glaucoma and retinal disorders
  • Face and Expression Recognition
  • Robotics and Sensor-Based Localization
  • Video Coding and Compression Technologies
  • Cryptography and Data Security
  • Digital Imaging for Blood Diseases
  • Cryptographic Implementations and Security
  • Advanced Vision and Imaging
  • Image and Signal Denoising Methods
  • Emotion and Mood Recognition
  • Advanced Image and Video Retrieval Techniques
  • Robotic Path Planning Algorithms
  • Handwritten Text Recognition Techniques
  • Cellular Automata and Applications

Jadavpur University
2016-2025

Ambedkar University Delhi
2020

Indian Statistical Institute
2015

Vascular map within the inner surface of retina provides insights about ophthalmic abnormalities or early signs different eye diseases. Devising a straight forward strategy to analyze vascular is quite challenging, given complex and delicate nature these vessels, as well high level noise in data. This article outlines new method for analyzing by extracting various features allocating weightage features. The most influential identifying have been determined using called Permutation...

10.36548/jiip.2024.4.006 article EN Journal of Innovative Image Processing 2025-01-01

This study introduces a multimodal sentiment analysis system to assess and recognize human pain sentiments within an Internet of Things (IoT)-enabled healthcare framework. integrates facial expressions speech-audio recordings evaluate intensity levels. integration aims enhance the recognition system’s performance enable more accurate assessment intensity. Such approach supports improved decision making in real-time patient care, addressing limitations inherent unimodal systems for measuring...

10.3390/s25041223 article EN cc-by Sensors 2025-02-17

In this work, we have proposed an audio encryption method. The method is signal sensitive as the hash value of given computed using SHA3-512, which returns a significantly large key size 512-bit. This used to set different parameters. work suggests 2D Cosine Logistic Map (2DCLM) by fusing map with map. 2DCLM functions well under chaos. scrambled help value. decomposed Empirical Mode Decomposition (EMD); before EMD segmented into reduce time complexity residuals and stream generated are...

10.1109/tla.2024.10472959 article EN IEEE Latin America Transactions 2024-03-14

Obstacle detection is an essential task for the autonomous navigation by robots. The becomes more complex in a dynamic and cluttered environment. In this context, RGB-D camera sensor one of most common devices that provides quick reasonable estimation environment form RGB depth images. This work proposes efficient obstacle tracking method using images to facilitate detection. To achieve early obstacles stable their states, as previous methods, we applied u-depth map Unlike existing present...

10.3390/s22176537 article EN cc-by Sensors 2022-08-30

Abstract This article proposes a multimodal sentiment analysis system for recognizing person’s aggressiveness in pain. The implementation has been divided into five components. first three steps are related to text-based perform classification tasks such as predicting the classes non-aggressive, covertly aggressive, and overtly aggressive classes. remaining two components an image-based system. A deep learning-based approach employed do feature learning predict types of pain An aggression...

10.1007/s12652-023-04567-z article EN cc-by Journal of Ambient Intelligence and Humanized Computing 2023-03-02

10.1007/s41315-023-00302-1 article EN International Journal of Intelligent Robotics and Applications 2023-11-14

Iris recognition has been paid more attentions due to its high reliability in personal identification recently. In this paper, an iris system proposed. The steps of the proposed method include localization, normalization, feature extraction and matching pattern. To describe data GLCM based Haralick features are used for purpose probabilistic neural network is employed. Experiments performed using images obtained from UBIRIS database. gives 97.00% correct classification rate.

10.1109/icectech.2011.5941793 article EN 2011-04-01

This paper presents a fast segmentation of iris portion from an eye image in recognition system. In segmentation, we have to find the inner boundary (between pupil and iris) outer sclera iris). To restricted circular Hough transform based method is applied. locate first passes through inversion then used. Both hough reduce search space Circular transform. The proposed faster one gives high accuracy result. performance tested with three standard databases MMU1, CASIA-Iris V3 IITD.

10.1109/icapr.2015.7050667 article EN 2015-01-01

10.1016/j.patcog.2004.02.008 article EN Pattern Recognition 2004-04-21

With the rapid growth in audio data volume, research area of content-based retrieval has gained impetus last decade. Audio classification serves as fundamental step towards it. Accuracy classifying relies on strength features and efficacy scheme. In this work, we have focused only. We restricted ourselves further time domain based low level features. Zero crossing rate (ZCR) shot energy (STE) are most widely used category. tried to develop reflecting quasi-periodic pattern signal by studying...

10.1109/iita.2009.427 article EN 2009-01-01

Iris localization is an important step in iris recognition system and performance of the depends on accuracy localization. In this paper, we have proposed method. Here, identify both outer inner circular boundary iris. The present method has two basic steps: edge points detection Hough transform. Before applying these steps, define bounding boxes for region pupil area. Bounding box reduces searching complexity transform results to a fast

10.1109/eait.2011.18 article EN 2011-02-01

10.1016/j.jvcir.2011.11.005 article EN Journal of Visual Communication and Image Representation 2011-11-19
Coming Soon ...