Saeid Hosseinipoor's Personal Page

I am a Ph.D. candidate in Computer Science at Stevens Institute of Technology focusing on computer vision. I have obtained my master's degree in Data Science and Analytics at University of Oklahoma.

My current research is mostly focused on depth estimation and dense label refining.

Dense Label Refiner

It is a deep learning projects that takes an estimated dense label and improves it.

Augmented Reality

This is a Augmented Reality project collaborating with Google. We use a Monocular Depth Estimator to find the depth insert an object in proper way to the scene.

Multiview 3D Reconstruction

This is a 3D Vision course project. We use multiple view of a scene and construct a point cloud of the object.

Dolly Zoom

This a 3D Vision course homework. We use a given point cloud to simulate the Dolly Zoom in cinematography by changing the camera's position and focal length simultaneously.

Segmentation

This is a Vision course homework. Segmentation methods (e.g. SLIC and SNIC) was implemented.

Natural Gas Price Prediction

This research thesis predicts the natural gas price in the US market by implementation of various methods including Neural Networks, Time Series, and Stochastic Processes.

I am interested in solving computer vision problems by geometric or deep learning methods. Specifically my interests are:
    - Binocular and stereo depth estimation
    - Binocular and multiple-view reconstruction
    - Semantic segmentation and object recognition
    - Augmented Reality (AR)
    - Deep convolutional network architecture design

Please leave me a message.