Paras Maharjan

PhD Student
Department of Computer and Electrical Engineering
University of Missouri - Kansas City, Missouri, USA
Supervisor: Dr. Zhu Li
Email: paras.maharjan@mail.umkc.edu

About Me

My name is Paras Maharjan. Currently, I am pursuing Ph.D. in Electrical and Computer Engineering at University of Missouri-Kansas City (UMKC). I completed MS in Electrical Engineering from same university in 2019 and BE in Electronics and Communication Engineering from the Tribhuvan University (TU), Kathmandu, Nepal in 2014. I am working as a Research Assistant in Multimedia Computing and Communication (MCC) Lab under the supervision of Professor Zhu Li.

Research Interest

Image Processing
Neural Network and Deep Learning
Computer Vision
Camera ISP
Image Enhancement
Internet of Things

Education

2020 - Present

PhD in Electrical Engineering

University of Missouri - Kansas City

Kansas City, Missouri, USA

2017 - 2019

MS in Electrical Engineering

University of Missouri - Kansas City

Kansas City, Missouri, USA

2010 - 2014

BE in Electronics and Communication Engineering

Tribhuvan University

Kathmandu, Nepal

Work Experience

2018 - Present

Graduate Research Assistant

Multimedia Computing and Communication (MCC) Lab

University of Missouri - Kansas City

Kansas City, MO, USA

May 2019 - Aug 2019

Camera Design Intern

Poly

Austin, TX, USA

2016 - 2017

Firmware Developer

Temco Controls Nepal

Lalitpur, Nepal

2014 - 2016

Firmware Developer

Real Time Solutions Pvt. Ltd

Lalitpur, Nepal

Publications

Paras Maharjan, Zhu Li, Ning Xu, Chongyang Ma, Li Li, Yue Li, "Improving Extreme Low-Light Image Denoising via Residual Learning", 2019 IEEE International Conference on Multimedia and Expo [pdf] [Web]

Research

DSFDNet [Web] [Code]

  • Proposed a nobel method of denoising before ISP
  • Decomposed the input raw image into low and high
    frequency subimages using wavelet transform
  • Design a new loss function for learning high frequency
    components of our proposed wavelet decomposition network

Low-light Image Enhancement [Web] [Code]

  • Proposed new end-to-end network for dark image enhancement
  • Implementated the residual learning and squeeze-and-excitation
    block for better performance gain

Occupancy map compression by LSTM and Arthmetic Coding

  • Prediction using LSTM
  • Code the sequence using arthmetic compression

Genome Sequence Compression

  • Extracted the coded region from human genome sequence
  • Prediction using LSTM
  • Code the sequence using arthmetic compression