Moktari Mostofa

PhD Candidate in Computer Science and Electrical Engineering

West Virginia University

About

Welcome to my homepage! I am a PhD Candidate in the Department of Computer Science and Electrical Engineering at West Virginia University , supervised by Dr. Nasser Nasrabadi.

Interests

Machine Learning, Deep learning, Computer vision, Object detection, Biometric Identification

Programming Skills

Platforms: PyTorch, TorchScript, Tensorflow, PyCharm, MATLAB and GIT.
Languages: Python, Matlab, C

Project Expertise

Face Recognition at Extreme Pitch and Yaw Angles:A novel coupled profile to the frontal PIFR model utilizing pose as an auxiliary information (i.e., pose attention) is developed.
Joint Pose Estimation and Face Frontalization:We have implemented a multi-task framework which is able to simultaneously rotate the off-angle face to a frontal view and gives an estimate of the pose angle of the input probe.
Benchmarking Video Super-Resolution Algorithm:In this project, we have investigated and benchmarked deep video super-resolution algorithms optimized for face recognition (FR), comprehensively examining the important factors in video super-resolution algorithms related to face recognition (quality, sharpness and clarity in facial features).
Cross-Spectral Iris Matching Using Coupled cGAN (conditional GAN): In this work, we have implemented a coupled generative adversarial network (CpGAN) architecture for cross-spectral iris recognition by projecting the visible and near-infrared iris images into a low-dimensional embedding domain to explore the hidden relationship between them. The framework investigates the potential capabilities of GAN based network to improve the performance of traditional cross-spectral iris recognition methods.
End-to-End Deep Super-Resolution Face Recognition System in the wild:We have designed a novel identity-aware GAN for face super resolution which enables us to hallucinate photo-realistic high-resolution (HR) faces while preserving the face identity. We combine the disciminator and face verifier by proposing a single network which performs both tasks simultaneously. Our discriminator jointly learns to distinguish face images at multiple scales. This unified multiscale structure enables the discriminator to transfer information between generated face images of different scales.
Joint Super Resolution and Vehicle Detection Network: We have designed a framework which implements super-resolution of low-resolution image and perform detection task on these super-resolved images simultaneously. The experimental results show that our proposed framework achieves better visual quality than the state-of-the-art methods for aerial super-resolution with 4x up-scaling factor and improves the accuracy of aerial vehicle detection.
Super Resolution-Assisted Deep Aerial Vehicle Detection: Our proposed system can be represented as a combination of two deep sub-networks. The first sub-network aims to use a GAN architecture for getting super resolved images. It ensures recovery of photo-realistic images from down-sampled images. The second sub-network consists of a deep neural network (DNN)-based object detector for detecting vehicles in super resolved images. In our architecture, the Single Shot Multi Box Detector (SSD) is used for vehicle detection.
Aerial Vehicle Detection on Super-resolved images using YOlOV3:We implement state-of-the-art object detection algorithm, YOLOV3 (You Only Look Once) and achieve satisfactory detection performance on super-resolved aerial images.