Prerana Mukherjee

Assistant Professor
Department of Computer Science
School of Engineering, JNU

Publications(Google Scholar, Google Scholar, Research Gate)
Teaching
Biography
Awards
Students
Contact Me

About Me

Welcome to my home page. I am part of School of Engineering, JNU working as Assistant Professor, Department of Computer Science. Prior to that, I was working as Assistant Professor at IIIT Sri CIty (Nov'18-Mar'20). During PhD'18 ( IIT Delhi ) I was working on research problem statement of "Salient Object Localization and Segmentation". I pursued M.Tech at Department of Information Technology, Delhi Technological University (DTU) , New Delhi. Passionate on working with bright and motivated students towards socio-economic and industry problems (healthcare, education, analytics and entertainment industries) in vision and allied areas. I have active collaborations with IBM, Intel, Accenture, Wipro and others. Open to industrial and academic collabortions.

Research Interests: Broad areas of vision (2D, RGBD & 3D point cloud based object localization & tracking, scene analysis, remote sensing and image/video processing), Learning Problems (Generative models, deep learning, sequential modelling and transfer learning), Biometrics, Pattern Recognition, Data Mining.

PhD and MS (research) programs:

I am actively looking for highly motivated research students both full-time (FT) and part time (PT) in the areas of image/vision processing, multimedia, NLP/AI. If you have commitment for research, critical thinking and analysis and zeal to do something for our nation, you are most welcome to join me. Interested students may contact me with a short statement of research they want to pursue and detailed CV.

M.Tech and B.Tech projects or internships (Research with me):

You may drop me an email and/or meet me directly in my office.

Curriculum Vitae Brief Bio-sketch

Latest Projects

Saliency based Image Segmentation

In this work, we present a novel approach for refining segmentation using saliency map. To achieve this, we first develop a new saliency detection method based on cues at various levels: local, global and rarity. Further, to increase the discriminability of the proposed saliency map, we apply a non-linear scale space preprocessor unit to the image. Finally the saliency map is used as initialization for segmentation. This technique is applied successively to perform multi-object segmentation.

Characterizing Objects

In this respect, we have developed a novel set of features using a combination of KAZE and SIFT keypoints termed as SIKA features. We have shown the efficacy of this feature set for object classification. SIKA features were designed to effectively characterize the objects. This is achieved by SIKA because these features serve as a good combination for boundary (KAZE) and saliency (SIFT) information. We have also shown that SIFT and KAZE are complementary features. Utilizing SIKA features, we develop a salient keypoint selection strategy based on the discriminability, detectability and repeatability of keypoints and show their effectiveness in object representation

Saliency and object feature driven segmentation

As, saliency map alone cannot always give the best segmentation because at times some background regions can be marked as salient. Also, the segmentation should closely fit the object contours to get neat segmentation. To overcome these issues, the technique is based on saliency and image features (Scale Invariant Feature Transform (SIFT), KAZE etc.) for characterization of objects and subsequent segmentation.

Ordered salient object proposals

We aim at demonstrating better performance (high recall and precision rates) utilizing very few number of object proposals. Thus, we are able to provide a tight bound on the number of object proposals. We rank the key objects in relative order of salience using the generated box proposals leading to lower complexity segmentation technique.


In Media

IITD Press Release on IoT Lab Establishment[Link1]. Navbharat Times on IoT Lab[Link2]. Bharti School Newsletter covering Celestini Award Ceremony[Link3]. Celestini Project India 2018 Award Ceremony by top media houses BusinessWire[Link4], Financial Express[Link5], Hindustan Times[Link6], Business Standard[Link7], India Today[Link8],NDTC Gadgets[Link9] and FirstPost[Link10]. Women's Day Coordination on theme AIML @IIIT Sri CIty[Link11]. Celestini Project India 2019[Link12].

Other Projects

Celestini Project India

We are working on increasing road safety by using video analytics from network of cameras mounted on vehicle dashboards in a more complex and chaotic environment like Indian roads particularly Delhi. My role involves mentoring and helping the student volunteers at IIT Delhi as well as development of algorithms required for the task.

Find out more

Github Code

View on GitHub

More on Facebook Page
                                                                                    

Sakshar Project

We would be working on an interactive platform designed to improve the learning by children through an interactive audio visual feedback system. This device uses a projector to render a virtual display, which permits production of large interactive displays.


                                                                                    
                                                                                    
1 / 3
Electric Circuitry
2 / 3
Prototype Model
3 / 3
Design Architecture

Teaching

1. Database Management System and Data Structures (Spring 2020)


SoE, Dept. of CSE, JNU
Role: Course Instructor.

2. Data Structures and Algorithms -I (Spring 2020)


Dept. of CSE, Computer Vision Group, IIIT Sri City
Role: Course Instructor.

3. Programming in C (Fall 2019)


Dept. of CSE, Computer Vision Group, IIIT Sri City
Role: Course Coordinator.

4. Computer Graphics and Multimedia (Fall 2019)


Dept. of CSE, Computer Vision Group, IIIT Sri City
Projects
Role: Course Instructor.

5. Database Management Systems (Spring 2019 & Fall 2019)


Dept. of CSE, Computer Vision Group, IIIT Sri City
Projects Course Page
Role: Course Instructor.

Students

PhD Students Guided


1. Chandra Sekhar Vorugunti, PhD persuing, IIIT Sri City (Co-guiding with Prof. P. Viswanath)
2. Jaswanth N., PhD persuing, IIIT Sri City (Collaborator)

B.Tech. Students Guided


1. Arvind Deshraj, IIIT Sri City
2. Aditya Haridas Menon, IIIT Sri City
3. Kaustubh Pandey, IIIT Sri City
4. Wasim Ishaq Khan, IIIT Sri City
5. Aakash Shukla, IIIT Sri City
6. Hareesh Thirukkovalluru, IIIT Sri City
7. Sri Harsha Varri, IIIT Sri City
8. Pavan Sai Mucherla, IIIT Sri City
9. G. Rusheel Kaushik, IIIT Sri City
10.Sagar, IIIT Sri City
11. Monica Sagar G., IIIT Sri City
12. B. Jahnavi, IIIT Sri City
13. N. Srilekha, IIIT Sri City
14. Parth Patwa, IIIT Sri CIty (B.Tech Honors)
15. Kevin John, IIIT Sri CIty (B.Tech Honors)
16. Rajat Singh, IIIT Sri CIty (B.Tech Honors)

Awards

1. Qualified UGC NET in December, 2012. Qualified GATE in February, 2011.
2. Received IEEE SPS Travel Grant for ICIP 2015 and ICIP 2017 Conference.
3. Received scholarship for securing second position during M.Tech. (Information Systems).
4. Received GATE (Graduate Aptitude Test in Engineering) scholarship [2011 - 2013].
5. Received IIT-Delhi (MHRD) institute fellowship [2013-2018].
6. Received Travel Grant from IIT Guwahati for presenting work at ICVGIP 2016 Conference.
7. Received IIT Delhi IRD Travel Grant of 1Lac for ICIP 2017 Conference.
8. 2nd position in ICPRAI 2018 conference competition.
9. Selected for MLSS 2018, Spain (150/500 candidates worldwide selected).
10. Shortlisted for NASA Frontier Development Lab 2018 Challenge.
11. Letter of Appreciation from Google for conducting TensorFlow Dev Summit Watch Party, 2019 at IIIT Sri City.
12. Winner Audio Visual Emotion Challenge (AVEC) 2019, ACM Multimedia Workshop, Nice France.
13. VisDrone MOT challenge 2019, ICCV Workshop, 8th Rank across the world on leaderboard.
14. VisDrone DET challenge 2019, ICCV Workshop, 38th Rank across the world on leaderboard.
15. Attended BRICS Young Scientist Conclave. Was shortlisted amongst 20 scientists over 450. applicants with a DST senior scientist representing INDIA globally.
16. Shortlisted for Final Round on Infosys Aarohan Social Innovation Awards to be held in Bangalore on 16th Jan, 2020.

Research

Journals

Characterizing objects with SIKA features for multiclass classification

Applied Soft Computing, Elsevier, 2016 (Impact Factor: 4.873)[PDF]

Saliency and KAZE features assisted object segmentation

Image and Vision Computing, Elsevier, 2017 (Impact Factor: 2.747)[PDF]

DeepFuseOSV: Online Signature Verification Using Hybrid Feature Fusion and Hybrid Depthwise Separable Convolution Neural Network Architecture

IET Biometrics, 2020 (Impact Factor: 2.092).[PDF]

Conditional Random Field based salient proposal set generation and its application in content aware seam carving

Signal Processing: Image Communication, Elsevier, 2020 (Impact Factor: 2.814)[PDF]

OSVFuseNet: Online Signature Verification by Feature Fusion and Depth-wise Separable Convolution Based Deep Learning.

Neurocomputing, Elsevier, 020 (Impact Factor: 4.072).[PDF]

Conferences

Point based features for contact-less palmprint images

2013 IEEE International Conference on Technologies for Homeland Security (HST)[PDF]

imPlag: Detecting image plagiarism using hierarchical near duplicate retrieval.

2015 Annual IEEE India Conference (INDICON)[PDF][PPT]

A real-time ball trajectory follower using Robot Operating System.

2015 IEEE Third International Conference on Image Information Processing (ICIIP)[PDF][PPT]

Automated ball tracking in tennis videos.

2015 IEEE Third International Conference on Image Information Processing (ICIIP)[PDF][PPT]

Adaptive Crypto-Steganosystem for videos based on Information Content and Visual Perception.

2015 IEEE Fifth National Conference on Computer Vision, Pattern Recognition, Image Processing and Graphics (NCVPRIPG)[PDF][POSTER][SPOTLIGHT PPT]

Saliency map based improved segmentation.

2015 IEEE International Conference on Image Processing (ICIP)[PDF][POSTER]

Adaptive image compression using saliency and KAZE features.

2016 International Conference on Signal Processing and Communications (SPCOM)[PDF][PPT]

Salient Keypoint Selection for Object Representation.

2016 IEEE Twenty Second National National Conference on Communications (NCC)[PDF][PPT]

SalProp: Salient object proposals via aggregated edge cues.

2017 IEEE International Conference on Image Processing (ICIP)[PDF][POSTER SIGPORT LINK]

Github Code

View on GitHub

Enhanced Characterness for Text Detection in the Wild.

2017 IAPR Computer Vision & Image Processing (CVIP-2017) and Workshop on Multimedia (WM-2017) [PDF] [PPT]

DRIZY- Collaborative Driver Assistance Over Wireless Networks.

2017 ACM The 23rd Annual International Conference on Mobile Computing and Networking (MobiCom 2017) [PDF][VIDEO][POSTER]

Object Classification using Ensemble of Local and Deep Features.

2017 IAPR International Conference on Advances in Pattern Recognition (ICAPR 2017) [PDF] [PPT]

Object cosegmentation using deep Siamese network.

2017 IEEE International Conference on Pattern Recognition and Artificial Intelligence (ICPRAI 2018) [PDF] [PPT]

Nrityantar: Pose oblivious Indian classical dance sequence classification.

2018 IAPR Indian Conference on Computer Vision, Graphics and Image Processing (ICVGIP)(ICVGIP 2018) [PDF][POSTER][SPOTLIGHT PPT]

OSVNet: Convolutional Siamese Network for Writer Independent Online Signature Verification.

2019 International Conference on Document Analysis and Recognition (ICDAR 2019) [PDF]

Online Signature Verification Based on Writer Specific Feature Selection and Fuzzy Similarity Measure.

2019 Applications of Computer Vision and Pattern Recognition to Media Forensics, CVPRW (CVPRW 2019) [PDF]

A Light weight and Hybrid Deep Learning Model based Online Signature Verification.

2019 The 2nd International Workshop on Machine Learning, (ICDAR-WML 2019) [PDF]

VisDrone-MOT2019: The Vision Meets Drone Multiple Object Tracking Challenge Results.

2019 VisDrone International Conference on Computer Vision Workshop (ICCVW 2019) [PDF]

Multi-level attention network using text, audio and video for Depression Prediction.

2019 Proceedings of the 9th International Workshop on Audio/Visual Emotion Challenge, AVEC 2019, ACM Multimedia Workshop (ACMMMW 2019) [PDF]][PPT]

Github Code

View on GitHub

VayuAnukulani: Adaptive Memory Networks for Air Pollution Forecasting.

2019 Global Conference on Signal and Information Processing (GlobalSIP) [PDF][PPT]

Github Code

View on GitHub

U-RME: Underwater Refined Motion Estimation in hazy, cluttered and dynamic environments.

2019 National Conference on Computer Vision, Pattern Recognition, Image Processing, and Graphics (NCVPRIPG 2019) [PDF][POSTER][SPOTLIGHT PPT]

Aerial Multi-Object Tracking by Detection Using Deep Association Networks.

2020 National Conference on Communications (NCC 2020) [PDF][PPT]

Prosody Features Characterization of Autism Speech for Automated Detection and Classification.

2020 National Conference on Communications (NCC 2020) [PDF][PPT]

Online Signature Profiling using Generative Adversarial Networks.

2020 International Conference on COMmunication Systems & NETworkS (COMSNETS)[PDF]

Aggression and Misogyny Detection using BERT: A Multi-Task Approach.

2020 European Language Resources Association (ELRA), Proceedings of the Second Workshop on Trolling, Aggression and Cyberbullying (TRAC-2020)[PDF] [PPT]

Github Code

View on GitHub

Countries Visited