Signal Processing

The Signal Processing research group primarily aims to design and develop efficient algorithms to resolve real-time and offline issues related to single and multi-dimensional signals/systems.

The research group works to develop innovative practices to upgrade the existing technologies as per the current industrial and social requirements.

Faculty members and students under this group work in different core and multidisciplinary research areas such as speech, audio and image processing, computer vision, and video processing, etc.

Research Focus

The Signal Processing research group focuses on data integration, techniques and tools necessary for efficient knowledge discovery to make proper decision-making processes in association with information extraction and computational intelligence.

In particular, the Signal Processing research team is working on the following.

Efficient 1-D filter design using evolutionary algorithms

Emotion recognition using speech signal analysis

Dark image enhancement under the presence of high intensity optical sources

2-D high frequency restoration using higher order derivative operators

Areas of Specialization

Signal and Speech Processing

Speech processing is a sub-area of digital signal processing used for the acquisition, manipulation, storage, transfer, analysis, enhancement, and recognition of speech signals. It has two primary aspects, namely speech recognition and speech synthesis. The Hidden Markov Model (HMM), Artificial Neural Network (ANN), and Dynamic Time Wrapping (DTW), etc. are a few of the recent techniques used in speech processing. Some of its futuristic applications include Interactive Voice Systems, Natural Language Processing, Virtual Assistants, Voice Identification, Emotion Recognition, Call Center Automation, etc.


Image and Video Processing

Image processing involves processing an image signal to improve its pictorial information for human interpretation and render it more suitable for autonomous machine perception. Recent research activities and scope in image processing include gesture analysis, enhancement, night vision, motion deblurring, computational photography, biomedical image processing, etc. Video processing is the extension of image processing in time-frame and falls under multidimensional signal processing. Video processing techniques are used in motion analysis, moving object detection, video compression, etc.


Computer Vision

Computer vision involves analysis, understanding, and recognition of human-centered visual data for machine perception and object recognition. It includes biometric recognitions such as the face, iris, fingerprint, vein pattern, palm, gait pattern, etc. Human emotions and nature too are well-perceived and understood through machine vision. The various areas related to emotion recognition include facial analysis, predicting identity, intent from faces, etc. In addition, computer vision has broad applicability in IoT, robotics, industrial automation, and agricultural sectors.


Biomedical Signal and Image Processing

Biomedical imaging concentrates on the capture and enhancement of images for both diagnostic and therapeutic purposes. Biomedical imaging technologies utilize either X-ray (CT scans), sound (ultrasound), magnetism (MRI), radioactive pharmaceuticals (nuclear medicine: SPECT, PET), or light (endoscopy, OCT) to assess the current condition of an organ or tissue. It can monitor a patient over time for diagnostic and treatment evaluation. It includes the analysis, enhancement, and display of images captured via x-ray, ultrasound, etc.


Featured Projects

07

Jan '2022

Signal Processing Research Group at SiliconTech implements a novel method for improving real time dark images for better night driving performance

Human vision is limited in perceiving objects under darkness, in the presence of highly illuminated light sources. Due to this, night-time driving becomes somewhat hazardous because the drivers see darker regions while oncoming vehicles are projecting their high beams into their eyes. During the night, potholes and sometimes road dividers, speed breakers, and humps are […]
READ MORE

06

Dec '2021

SiliconTech receives grant to carry out research work under TARE - SERB scheme, Government of India

Dr. Biranchi Narayan Rath, received a grant of ₹15 lakhs under the Teachers Associateship for Research Excellence (TARE), Science and Engineering Research Board (SERB), Government of India to carry out research in the area of active noise control. TARE scheme aims to facilitate faculty members from state universities, colleges, and private educational institutions to conduct research […]
READ MORE

Projects, People & Publications

This project aims at developing android apps for automatic emotion analysis from the voice data. Here we have created a database in Odia language for emotion analysis.

Understanding one’s feelings at the time of communication is constructive in comprehending the conversation and responding appropriately. Currently, this part of human–computer interaction has not yet entirely been solved, and except for a limited number of applications, there is no general solution to this problem.

In recent years in the context of the rapid development of Artificial Intelligence, speech emotion recognition is gradually becoming a challenging research. Human language communication has two modes, written mode and spoken mode. Though linguistic information is the same in both the modes, the latter also carries some paralinguistic information.

The paralinguistic information refers to the speech sounds that carry other speaker related information like, age, sex, emotion and attitude .The most important and complex of them is emotion.

Recognizing emotional content in speech is an indispensable requirement for efficient human-machine interaction.

Human vision is limited to perceive the objects under darkness in presence of highly illuminated light sources. This causes a major problem during night driving by the drivers to watch at darker regions when high beams of vehicle headlights are projected on their eyes.

The same problem arises while capturing information by the traffic management system due to direct projection of street light and vehicle high beam on a CCD camera. Very little research has been carried out to address this problem but no significant results have come out till now.

The existing dark image enhancement algorithms can effectively enhance the darker regions but fail to enhance the brighter regions that are over saturated. This leads to fading of the nearby darker objects resulting in information loss.

Hence, there is a requirement to develop an algorithm which can not only enhance the dark regions of an image but at the same time effectively manipulate the local contrast of the bright regions for proper overall enhancement of an image without any significant loss of original information.

Potholes and speed breakers on road can generate severe damage to tires and vehicles. Especially in India, vehicle collision and accidents happen due to immense depth of potholes on road. Detection and repairing of potholes and speed breakers is a major challenge for Intelligent Transportation System (ITS) service and road management system.

The proposed algorithm not only detects potholes and speed breaker in real-time but also gives an enhanced, highlighted video output for proper recognition of potholes and speed breakers under varieties of circumstances.

The algorithm makes use of a 2-D high pass filter for detecting the outlines of potholes and speed breakers since they represent high frequency. A newly developed horizontal, vertical and diagonal (HVD) filter is developed which enhances the variations in horizontal, vertical and diagonal directions to highlight the details of speed breakers and potholes effectively. Subsequently, the detected region is segmented and filled with colour so that the region will be properly highlighted and visible to the driver in real-time.

Electronics and Communication Engineering

Electronics and Communication Engineering

Electronics and Communication Engineering

Electronics and Communication Engineering

Electronics and Communication Engineering

Electronics and Communication Engineering

Electronics and Communication Engineering

Computer Applications

Dept. of Electronics & Instrumentation Engineering, SiliconTech

Dept. of Electronics & Communication Engineering, SiliconTech

B. Tech. Electronics and Communication Engineering, 2017-2021

B. Tech. Electronics and Communication Engineering, 2017-2021

B. Tech. Electronics and Communication Engineering, 2017-2021

B. Tech. Electronics and Communication Engineering, 2017-2021

B. Tech. Electronics and Communication Engineering, 2017-2021

Indian Institute of Engineering Science and Technology Shibpur, Sep 2020 - Present, 2021

DRDO Defence Institute, Oct 2019 - Jul 2020

IIT Dhanbad, Aug - Sept 2019

Judhisthir Dash, Bivas Dam, and Rajkishore Swain

, "Design and implementation of sharp edge FIR filters using hybrid differential evolution particle swarm optimization
", International Journal of Electronics and
Communications (AEU)
, Elsevier, Vol. 114, pp. 1-16, 1123,

DOI: 10.1016/j.aeue.2019.153019

.

Nalini Singh and Satchidananda Dehuri

, "Multiclass classification of EEG signal for epilepsy detection using DWT based SVD and fuzzy kNN classifier ", Intelligent Decision Technologies-An International Journal, Vol. 14, No. 2, pp. 239-252, 0601,

DOI: 10.3233/IDT-190043

.

Judhisthir Dash, Bivas Dam, and Rajkishore Swain

, "Improved firefly algorithm based optimal design of special signal blocking IIR filters", Measurement, Elsevier, Vol. 149, Article ID: 106986, 0823,

DOI: 10.1016/j.measurement.2019.106986

.

Annapurna Mishra and Satchidananda Dehuri

, "Real-time online fingerprint image classification using adaptive hybrid techniques", International Journal of Electrical & Computer Engineering, Vol. 9, No. 5, 1023,

DOI: 10.11591/ijece.v9i5.pp4372-4381

.

Manoroma swain, A. Routray, and P. Kabisatapathy

, "Databases features & classifiers – A review", International Journal of Speech Technology, Springer, Vol. 21, No. 1, pp. 93-120, 0323,

DOI: 10.1007/s10772-018-9491-z

.

Bhagyalaxmi
Jena, A. Jena, and
S.S. Singh.

, "Stress speech recognition using support vector machine and random forest", International
journal of
public health
research and
development
, Vol. 9, No. 9, pp. 1014-1018, 1023,

DOI: 10.5958/0976-5506.2018.01134.8

.

Bhagyalaxmi Jena and Sudhansu Sekhar Singh

, "An approach to spectral analysis of psychologically influenced speech", International Journal of Engineering & Technology, Vol. 7, No. 1.2, pp. 66-70, 1223.

Bhagyalaxmi Jena and Sudhansu Sekhar Singh

, "Analysis of stressed speech on teager energy operator (TEO)", International Journal of Pure & Applied Mathematics,, Vol. 118, No. 16, pp. 667-680, 0623.

Abhilash Mohanty, Bhagyalaxmi Jena, Sruti Nanda, and S. S. Singh

, "Psychological Stress Speech Analysis:", Asian Journal of Convergence in Technology, Vol. III, Vol. III, pp. 14-20, 0623.

Aditya Acharya and Sukadev Meher

, "Efficient fuzzy composite predictive scheme for effectual 2-D up-sampling of images for multimedia applications. ", Journal of visual communication & Image representation, Elsevier, Vol. 44, pp. 156-186, 0423,

DOI: 10.1016/j.jvcir.2017.01.031

.

Aditya Acharya and Sukadev Meher

, "Composite high frequency predictive scheme for efficient 2-D up-scaling performance", Multimedia tools & applications, Springer, Vol. 76, No. 1, pp. 1-37, 0123,

DOI: 10.1007/s11042-017-4346-1

.

Annapurna Mishra, Satchidananda Dehuri, Pradeep kumar Mallick

, "Classification of Real Time Noisy Fingerprint Images Using FLANN", in Proceedings of International Conference on Biologically Inspired Techniques in Many-Criteria Decision Making , Springer, pp.62-68, 0123,

DOI:
10.1007/978-3-030-39033-4_6

.

Annapurna Mishra and Satchidananda Dehuri

, "Fingerprint Classification by Filter Bank Approach Using Evolutionary ANN", in Proceedings of CISC 2017, Cognitive Informatics and Soft Computing, Springer, VBIT, Hyderabad, pp. 343-351, 0823,

DOI: 10.1007/978-981-13-0617-4_34

.

Monorama Swain, Aurobinda Routray, P. Kabisatpathy, and J. N. Kundu

, "Study of Prosodic Feature Extraction for Multidilectal Odia Speech Emotion Recognition", in Proceedings of IEEE Tencon region '10, IEEE, Singapore, pp. 1644-1649, 0623,

DOI:
10.1109/TENCON.2016.7848296

.

Explore Silicon’s research in diverse areas developing technological innovations to advance society.

View Research Domains