Dr. Pratik Chattopadhyay

Assistant Professor
Department/School/Unit Name
Department of Computer Science and Engineering
Phone No(s): 0542-716-5325
Email: pratik.cse@iitbhu.ac.in
Area of Interest: Image and Video Processing, Computer Vision, Pattern Recognition, Machine/Deep Learning, Generative Modelling

I am working as an Assistant Professor at the Department of Computer Science & Engineering at the Indian Institute of Technology (Banaras Hindu University), Varanasi. I am also serving as an Associate Editor of the IEEE Transactions on Systems, Man, and Cybernetics: Systems.

Prior to this, I did my post-doctoral studies for one year at the School of Electrical & Electronic Engineering, Nanyang Technological University, Singapore, and for six months at the Electronics and Communication Sciences Unit, Indian Statistical Institute, Kolkata, India.

I completed my Ph.D. from the Indian Institute of Technology, Kharagpur, India in 2015. The broad area of my Ph.D. research was Image and Video Processing.

I received Master degree (MCSE) in Computer Science & Engineering from Jadavpur University, Kolkata, India in 2011, and B Tech degree in Computer Science & Engineering from West Bengal University of Technology, India in 2009.

My present research interests include the application of Machine/Deep Learning in Computer Vision and Image Processing, Cyber-Security, Transporation, and other domains.

My hobbies are painting, handicraft, and playing musical instruments.

Even Semester (2023-24)
Artificial Intelligence (CSE 241)
Selected Topics in Artificial Intelligence (CSE 503) (Course shared with Dr. Bidyut Patra)
Data Structures (CSO 102)

Odd Semester (2023-24)
Intelligent Systems(CSE 342)
Intelligent Systems (CSE 523)

Even Semester (2022-23)
Artificial Intelligence (CSE 241)
Selected Topics in Artificial Intelligence (CSE 503) (Course shared with Dr. Kailasam Lakshmanan)
Data Structures (CSO 102)

Odd Semester (2022-23)
Intelligent Computing (CSE 342)
Intelligent Systems (CSE 523)

Even Semester (2021-22)
Artificial Intelligence (CSE 241)
Selected Topics in Artificial Intelligence (CSE 503) (Course shared with Dr. Kailasam Lakshmanan)
Selected Topics in Machine Learning (CSE 540) (Course shared with Dr. Amrita Chaturvedi and Dr. Kailasam Lakshmanan)

Odd Semester (2021-22)
Intelligent Computing (CSE 342)

Even Semester (2020-2021)
Image Processing & Computer Vision (CS 7006)
Pattern Recognition (CS 7010)
Advanced Computational Mathematics (CS 7008)
Data Structures (CSO 102)

Odd Semester (2020-21)
Computer Graphics (CSO 351)
Even Semester (2019-2020)
Computer Programming (CSO 101)
Machine Learning (CSE 464)
Odd Semester (2019-20)
Computer Programming (CSO 101)
Intelligent Computing (CSE 342)
Even Semester (2018-2019)
Computer Programming (CSO 101)
Machine Learning (CSE 464)
Biometrics (CS 7005)

Odd Semester (2018-2019)
Computer Programming (CSO 101)

Even Semester (2017-2018)
Computer Programming (CSO 101)
Biometrics (CS 7005)

Odd Semester (2017-2018)
Computer Programming (CSO 101)


Binit Singh, Divij Singh, Rohan Kaushal, Agrya Halder, and Pratik Chattopadhyay, "GSSTU: Generative Spatial Self-Attention Transformer Unit for Enhanced Video Prediction", IEEE Transactions on Neural Networks and Learning Systems (Accepted in 2024).

Somnath Sendhil Kumar, Binit Singh, Pratik Chattopadhyay, Agrya Halder, and Lipo Wang, BGaitR-Net: An Effective Neural Model for Occlusion Reconstruction in Gait Sequences by Exploiting the Key Pose Information, Expert Systems with Applications (Accepted in 2024).

Harshit Agrawal, Agrya Halder, and Pratik Chattopadhyay, A Systematic Survey on Recent Deep Learning-based Approaches to Multi-Object Tracking, Multimedia Tools and Applications (Accepted in 2023).

Sanjay Kumar Gupta and Pratik Chattopadhyay, Pose-Based Boundary Energy Image for Gait Recognition from Silhouette Contours, Sadhana (Accepted in 2023).

Nirbhay Kumar Tagore, Prathistith Raj Medi, and Pratik Chattopadhyay, Deep Pixel Regeneration for Occlusion Reconstruction in Person Re-identification, Multimedia Tools and Applications (Accepted in 2023).

Preetam Pal, Pratik Chattopadhyay, and Mayank Swarnkar, Temporal Feature Aggregation with Attention for Insider Threat Detection from Activity Logs, Expert Systems with Applications (Accepted in 2023).

Utkarsh Mishra, Akshat Agrawal, Josephine Crystal R Mathew, Rajesh Kumar Pandey, and Pratik Chattopadhyay, An Efficient Approach for Image De-fencing based on Conditional Generative Adversarial Network, Signal, Image and Video Processing 17 (1), 147-155, 2023.

Agrya Halder, Pratik Chattopadhyay, and Sathish Kumar, Gait Transformation Network for Gait De-Identification with Pose Preservation, Signal, Image and Video Processing, 10.1007/s11760-022-02386-x, 2022. (Accepted)

Rajarajeswari Perepi, Pratik Chattopadhyay, and Anwar Bég O, A Deep Learning Computational Approach for the Classification of COVID-19 Virus, Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 10.1080/21681163.2022.2111722, 2022. (Accepted)

Nirbhay Kumar Tagore, and Pratik Chattopadhyay. A Bi-Network Architecture for Occlusion Handling in Person Re-Identification, Signal, Image and Video Processing, 16,1071-1079, 2022.

Sanjay Kumar Gupta and Pratik Chattopadhyay, Gait Recognition in the Presence of Covariate Conditions, Elsevier Neurocomputing, 454: 76-87, 2021.

Nirbhay Kumar Tagore, Ayushman Singh, Manche Sumanth, Pratik Chattopadhyay, Person Re-identification from Appearance Cues and Deep Siamese Features, Journal of Visual Communication and Image Representation, 75, Art. No. 103029, 2021.

Ayush Agarwal, Pratik Chattopadhyay, Lipo Wang, Privacy Preservation through Facial De-identification with Simultaneous Emotion Preservation, Signal, Image and Video Processing, 15, 951-958, 2021.

Sanjay Kumar Gupta and Pratik Chattopadhyay, Exploiting Pose Dynamics for Human Recognition from their Gait Signatures, Multimedia Tools and Applications, 80, 35903-35921, 2021.

Divyanshu Gupta, Shorya Jain, Utkarsh Tripathi, and Lipo Wang, A Robust and Efficient Image De-fencing Approach using Conditional Generative Adversarial Networks, Signal, Image and Video Processing, 15, 297-305, 2021.

Nirbhay Kumar, Pratik Chattopadhyay, Lipo Wang, T-MAN: A Neural Ensemble Approach for Person Re-Identification using Spatio-temporal Information, Multimedia Tools and Applications, 79(37), 28393–28409, 2020.

Pratik Chattopadhyay, Lipo Wang, Yap-Peng Tan, Scenario-Based Insider Threat Detection From Cyber Activities, IEEE Transactions on Computational Social Systems, 5(3), 1-16, 2018. 

Aditi Roy, Pratik Chattopadhyay, Shamik Sural, Jayanta Mukherjee and Gerhard Rigoll, Modeling, Synthesis and Characterization of Occlusion in Videos, IET Computer Vision, 9(6): 821-830, 2015.

Pratik Chattopadhyay, Shamik Sural, and Jayanta Mukherjee, Information Fusion from Multiple Cameras for Gait-based Re-identification and Recognition, IET Image Processing, 9(11): 969-976, 2015.

Pratik Chattopadhyay, Shamik Sural, and Jayanta Mukherjee, Frontal Gait Recognition from Occluded Scenes, Elsevier Pattern Recognition Letters, 63(1): 9-15, 2015. 

Pratik Chattopadhyay, Shamik Sural, and Jayanta Mukherjee, Frontal Gait Recognition from Incomplete Sequences using RGB-D Camera, IEEE Transactions on Information Forensics and Security, 9(11): 1843-1856, 2014.

Pratik Chattopadhyay, Aditi Roy, Shamik Sural, and Jayanta Mukhopadhyay, Pose Depth Volume Extraction from RGB-D Streams for Frontal Gait Recognition, Elsevier Journal of Visual Communication and Image Representation, 25(1): 53–63, 2014.


Dhruv Gupta, Dhruve Kiyawat, V Venkata Vinay Kumar, Utkarsh Mishra and Pratik Chattopadhyay, DefenceLite: An Effective Lightweight GAN-based Image De-fencing Model, 9th International Congress on Information and Communication Technology, 2024. (Accepted in 2024).

Harshit Agrawal, Agrya Halder, and Pratik Chattopadhyay, MotionFormer: An Improved Transformer-Based Architecture for Multi-Object Tracking, 8th International Conference on Computer Vision & Image Processing (Accepted in 2023).

Dhritimaan Das, Ayush Agarwal, Pratik Chattopadhyay, Gait Recognition from Occluded Sequences in Surveillance Sites, Workshop on Real-World Surveillance in conjunction with European Conference on Computer Vision, Springer Nature Switzerland, 703-719, 2022.

Ankit Sinha, Soham Banerjee, Pratik Chattopadhyay, Effective Stacking of Deep Neural Models for Automated Object Recognition in Retail Stores, International Conference on Image Analysis and Processing, (Accepted in 2022).

Alakh Aggarwal, Rishika Rathore, Pratik Chattopadhyay, Lipo Wang, EPD-Net: A GAN-based Architecture for Face De-identification from Images, IEEE International Conference on IOT, Electronics and Mechatronics Conference (IEMTRONICS), 1-7, 2020.

Nirbhay Tagore, Pratik Chattopadhyay SMSNet: A Novel Multi-Scale Siamese Model for Person Re-Identification, 17th International Conference on Signal Processing and Multimedia Applications. (Accepted in 2020).

Krishnan, Utsav, Ayush Agarwal, Avinash Senthil, and Pratik Chattopadhyay. Image Enhancement and Denoising in Extreme Low-Light Conditions. (Accepted in 2019).

Shailesh Shrivastava, Alakh Aggarwal, Pratik Chattopadhyay, Broad Neural Network for Change Detection in Aerial Images, 1st International Conference on Emerging Technology in Modelling and Graphics, Springer Singapore, 327-339, 2019.

Sanjay Kumar Gupta, Gaurav Mahesh Sultaniya, Pratik Chattopadhyay, An Efficient Descriptor for Gait Recognition using Spatio-Temporal Cues, International Conference on Emerging Technology in Modelling and Graphics, Springer Singapore, 85-97, 2019.

Utsav Krishnan, Akshal Sharma, Pratik Chattopadhyay, Feature Fusion from Multiple Paintings for Generalized Artistic Style Transfer, International Conference on Advances in Engineering Science Management & Technology, SSRN 3387817, 2019. 

David Lorenzi, Pratik Chattopadhyay, Emre Uzun, Jaideep Vaidya, Shamik Sural, Vijayalakshmi Atluri, Generating Secure Images for CAPTCHAs through Noise Addition, ACM Symposium on Access Control Models and Technologies, 169-172, 2015. (demo paper)

Pratik Chattopadhyay, Shamik Sural and Jayanta Mukherjee, Exploiting Pose Information for Gait Recognition from Depth Streams, 4th IEEE Workshop on Consumer Depth Cameras for Computer Vision in conjunction with European Conference on Computer Vision, Lecture Notes in Computer Science, Springer International Publishing, 341-355, 2014.

Pratik Chattopadhyay, Shamik Sural, and Jayanta Mukherjee, Gait Recognition from Front and Back View Sequences captured using Kinect, 5th International Conference on Pattern Recognition and Machine Intelligence, Lecture Notes in Computer Science, Springer Berlin Heidelberg, 196–203, 2013.


Pratik Chattopadhyay, Shamik Sural, Jayanta Mukhopadhyay, Surveillance using Partial Gait Sequences, Status: Published

Book Chapter

Ashish Kumar, Sachin Srivastava, Pratik Chattopadhyay, Machine and Deep‐Learning Techniques for Image Super‐Resolution in Machine Learning Algorithms for Signal and Image Processing, John Wiley & Sons, Inc., 89-113.


    S. No.

    Title of Project

    Funding Agency






    Developing Improved Algorithms for Intelligent Video Surveillance

    Core Research Grant, Science and Engineering Research Board, Dept. of Science & Technology, India

    INR 28,09,400/- 



    In Progress


    Development of a Lightweight Android Mobile Software Powered by Deep Learning for Identification of Plant Leaf Disease

    Technology Development Programme, Dept. of Science & Technology, India

    INR 35,94,000/-



    In Progress

    3 Development of a Disaster Response System for Collecting and Disseminating Information through Social-Media Text Processing Uttar Pradesh: Council of Science and Technology - 2022-2025 co-PI In Progress


    Lab Server 1: Two graphics processing units (GPUs): Nvidia Titan Xp with 12-GB RAM, total FB memory as 12196 MB and total BAR1 memory as 256 MB, and a Nvidia GeForce GTX 1080 Ti with 11-GB RAM, total FB memory as 11178 MB and total BAR1 memory as 256 MB, OS: CentOS, System RAM 96 GB.

    Lab Server 2: Dell Precision 7920 Tower Workstation, 40x IntelXeon Gold 5215 CPU@2.50GHz, 93 GB RAM, NVIDIA GeForce RTX 3090/PCIe/SSE2.

    Research Theme: Computer Vision, Pattern Recognition, Machine/Deep Learning Applications


    Main Research Activities Include:

    Person Re-identification

    Person re-identification plays a central role in tracking and monitoring crowd movement in public places, and hence it serves as an important means for providing public security in video surveillance application sites. The problem of person re-identification has received significant attention in the past few years, and with the introduction of deep learning, several interesting approaches have been developed. In one approach, we have developed an ensemble model called Temporal Motion Aware Network (T-MAN) for handling the visual context and spatio-temporal information jointly from the input video sequences. Our methodology makes use of the long-range motion context with recurrent information for establishing correspondences among multiple cameras. The proposed T-MAN approach first extracts explicit frame-level feature descriptors from a given video sequence by using three different sub-networks (FPAN, MPN, and LSTM), and then aggregates these models using an ensemble technique to perform re-identification. The method has been evaluated on three publicly available data sets, namely, the PRID-2011, iLIDS-VID, and MARS, and re-identification accuracy of 83.0%, 73.5%, and 83.3% have been obtained from these three data sets, respectively. 

    We are also working on other dimensions of the problem and challenges such as occlusion and open-set re-identification.

    Gait Recognition

    Computer vision-based gait recognition has evolved into an active area of research since the past decade, and a number of useful algorithms have been proposed over the years. Among the existing gait recognition techniques, pose-based approaches have gained more popularity due to their inherent capability of capturing the silhouette shape variation during walking at a high resolution. However, a short-coming of the existing pose-based gait recognition approaches is that their effectiveness depends on the accuracy of a pre-defined set of key poses and are, in general, not robust against varying walking speeds. We have proposed an improvement to the existing pose-based approaches by considering a gallery of key pose sets corresponding to varying walking speeds instead of just a single key pose set. This gallery is generic and is constructed from a large set of subjects that may/may not include the subjects present in the gait recognition data set. Comparison between a pair of training and test sequences is done by mapping each of these into the individual key pose sets present in the above gallery set, computing the Active Energy Image for each key pose, and next observing the frequency of matched key poses in all the sets. Our approach has been evaluated on two popular gait data sets, namely the CASIA B data and the TUMGAID data.

    We are also working on handling occlusion in gait recognition and view-invariant gait recognition.

    Face and Gait De-identification

    Due to the availability of low-cost internet and other data transmission media, a high volume of multimedia data gets shared very quickly. Often, the identity of individuals gets revealed through images or videos without their consent, which affects their privacy. Since face is the only biometric feature that reveals the most identifiable characteristics of a person in an image or a video frame, the need for the development of effective face and gait de-identification algorithms for privacy preservation cannot be over-emphasized. Existing solutions to face de-identification are either non-formal or are unable to obfuscate identifiable features completely. We are working towards developing automated face de-identification algorithms that take as input a facial image and generate a new face preserving the emotion and non-biometric facial attributes of a target face. Also, there has been no work on gait de-identification to date. Unlike face de-identification, it requires processing of a set of frames corresponding to a complete gait cycle. Generative Modelling techniques are being employed to carry out this work.

    Insider Threat Detection

    An insider threat scenario refers to the outcome of a set of malicious activities caused by intentional or unintentional misuse of the organization's systems, networks, data, and resources. Prevention of insider threat is difficult, since trusted partners of the organization are involved in it, who have authorized access to these confidential/sensitive resources. The state-of-the-art research on insider threat detection mostly focuses on developing unsupervised behavioral anomaly detection techniques with the objective of finding out anomalousness or abnormal changes in user behavior over time. However, an anomalous activity is not necessarily malicious that can lead to an insider threat scenario. As an improvement to the existing approaches, we propose a technique for insider threat detection from time-series classification of user activities. Initially, a set of single-day features is computed from the user activity logs. A time-series feature vector is next constructed from the statistics of each single-day feature over a period of time. The label of each time-series feature vector (whether malicious or non-malicious) is extracted from the ground truth. To classify the imbalanced ground-truth insider threat data consisting of only a small number of malicious instances, we employ a cost-sensitive data adjustment technique that under-samples the non-malicious class instances randomly. As a classifier, we employ a two-layered deep autoencoder neural network and compare its performance with other popularly used classifiers: random forest and multilayer perceptron. Encouraging results are obtained by evaluating our approach using the CMU Insider Threat Data, which is the only publicly available insider threat data set consisting of about 14-GB web-browsing logs, along with logon, device connection, file transfer, and e-mail log files.

    Image Defencing

    Image de-fencing is one of the most important aspects of recreational photography in which the objective is to remove the fence texture present in an image and generate an aesthetically pleasing version of the same image without the fence texture. We are developing automated and effective techniques for fence removal and image reconstruction using conditional generative adversarial networks (cGANs). These networks have been successfully applied in several other domains of computer vision, focusing on image generation and rendering. One of our approaches is based on a two-stage architecture involving two cGANs in succession, in which the first cGAN generates the fence mask from an input fenced image, and the next one generates the final de-fenced image from the given input and the corresponding fence mask obtained from the previous cGAN.

    Image Denoising

    Image noise refers to the specks of false colors or artifacts that diminish the visual quality of the captured image. It has become our daily experience that with affordable smart-phone cameras we can capture high clarity photos in a brightly illuminated scene. But using the same camera in a poorly lit environment with high ISO settings results in images that are noisy with irrelevant specks of colors. Noise removal and contrast enhancement in images have been extensively studied by researchers over the past few decades. But most of these techniques fail to perform satisfactorily if the images are captured in an extremely dark environment. In recent years, computer vision researchers have started developing neural network-based algorithms to perform automated de-noising of images captured in a low-light environment. Although these methods are reasonably successful in providing the desired de-noised image, the transformation operation tends to distort the structure of the image contents to a certain extent. We propose an improved algorithm for image enhancement and de-noising using the camera’s raw image data by employing a deep U-Net generator. The network is trained in an end-to-end manner on a large training set with suitable loss functions. To preserve the image content structures at a higher resolution compared to the existing approaches, we make use of an edge loss term in addition to PSNR loss and structural similarity loss during the training phase.

    Current Ph.D. Students

    1. Preetam Pal: Topic: Insider Threat Detection
    2. Anshita Malviya: Topic: Semi-supervised Learning in Medical Image Segmentation
    3. Harsh Raj Singh: Topic TBD

    Past Ph.D. Students

    1. Sanjay Kumar Gupta: Thesis Defended (Title: Developing Improved Algorithms for Pose-Based Gait Recognition)
    2. Dr. Nirbhay Kumar Tagore: (Title: Person Re-identification for Intelligent Surveillance using Deep Learning) Currently Assistant Professor at Rajiv Gandhi Institute of Petroleum Technology

    ​​Conference Photos

    My Paintings

    My Musical Instruments

    Email: pratik . cse @ iitbhu . ac . in / pratikchattopadhyay @ gmail . com

    Facebook Profile

    Youtube Channel