Sina M. Baharlou
Research Assistant at Boston University
Created by Sina M. Baharlou
Ph.D. in Electrical and Computer Engineering
Boston University, Massachusetts
January 2023
major: Machine Learning for Photonics
adviser: Prof. Abdoulaye Ndao
M.Sc. in Artificial Intelligence and Robotics
Sapienza University of Rome, Italy
September 2014 — February 2017
thesis: Depth Colorization using Deep CNNs
major: Visual Learning
adviser: Prof. Barbara Caputo
grade: 110/110
GPA: 28/30
B.Sc. in Computer Software Engineering
Islamic Azad University, South Tehran Branch, Iran
September 2009 — February 2014
thesis: Automatic License Plate Recognition Software for Persian License Plates
major: Image Processing
adviser: Prof. Shahrdad Ezzatdoust
GPA: 15.44/20
HSD in Mathematics and Physics
Dr. Hessabi High School, Tehran, Iran
September 2005 — June 2008
GPA: 18.42/20
Research Assistant
Boston University
Massachusetts, USA
January 2023 — Present
Description: I am working on utilizing Machine Learning for Photonics applications under the supervision of Prof. Ndao.
Visiting Researcher
Ludwig Maximilian University of Munich
Tehran, Iran
July 2019 — December 2021
Description: I collaborated on several projects with a research team from the Ludwig Maximilians University of Munich under the supervision of Prof. Volker Tresp. During this time, I closely worked with Dr. Sahand Sharifzadeh on the topics of Scene Graph Classification and Visual Relation Detection. The main focus of our research was to model prior knowledge and utilize it in scene understanding. We investigated various sources of prior knowledge including depth map image pairs, raw textual data, and curated knowledge graphs. During these projects we also explored some relevant topics including Knowledge Graph Embedding models and Language Models for Link Prediction. The outcome of our works is three scientific papers published in International Conference on Pattern Recognition and AAAI Conference on Artificial Intelligence.
Research Experience on Transfer Learning and Vehicle Recognition
Independent Research
Tehran, Iran
June 2017 — March 2018
Description: I conducted a research project with a comparative approach aimed to determine the robustness of deep neural networks trained on a noisy dataset using the transfer learning technique. The proposed method was applied in a vehicle recognition scenario and resulted in a publication in the journal of Expert Systems with Applications. This project was done in collaboration with Prof. Abbasi Aghamaleki.
Research Collaboration on Deep Depth Colorization
Sapienza University of Rome
Rome, Italy
November 2016 — March 2017
Description: I contributed to the early research on 'Depth Colorization' where we proposed a deep learning approach of normalizing and mapping the depth images into three channels in order to improve the recognition accuracy of RGB-D images (Nov. 2016 - Mar. 2017). I Also assisted in the preparation phase of a synthetic depth dataset called 'VANDAL' which is used to train ConvNets (Feb. 2016 - Apr. 2016).