I am a MAsc student at the Human Communication Technologies Lab. My research interest is in 3D computer vision including but not limited to Multiple View Geometry, Structure from Motion, Multi-View Stereo, Photometric Stereo, Structured Light, Shape from X (Shading, Silhouette), camera calibration, camera-projector calibration.

CS+MATH GPA: 3.92/4, Cumulative GPA: 3.7/4
Thesis: Development and Evaluation of A 3D Reconstruction Framework for General Objects

GPA: 86.4/100 or 3.6/4

Qian Zhou, Kai Wu, Gregor Miller, Ian Stavness and Sidney Fels, "3DPS: An auto-calibrated three-dimensional perspective-corrected spherical display" IEEE Virtual Reality (IEEEVR), 2017

Q. Zhou, G. Miller, K. Wu, S. Fels. "Automatic Calibration of a Multiple-Projector Spherical Fish Tank VR Display." Winter Conference on Applications of Computer Vision, 2017.

Q. Zhou, G. Miller, K. Wu, I. Stavness, and S. Fels. "Analysis and practical minimization of registration error in a spherical fish tank virtual reality system." Asian Conference on Computer Vision. Springer, 2016.

3D Computer Vision: Multiple View Geometry, Structure from Motion, Multi-View Stereo, Photometric Stereo, Structured Light, Shape from Silhouette

• Created the datasets to evaluate the performance of algorithms under varied problem conditions (visual/geometric properties) so that the conditions under which a specific algorithm works reliably can be discovered;
• Created the datasets to evaluate the performance of algorithms under varied problem conditions (visual/geometric properties) so that the conditions under which a specific algorithm works reliably can be discovered;
• Developed Blender projects and scripts to generate synthetic datasets with varied material and geometric properties, which make it possible to discover the problem conditions under which a specific algorithm works reliably.

• Developed a description-based interface to 3D reconstruction so that a successful reconstruction result can be achieved given a valid description of the object;
• Implemented multiple 3D reconstruction algorithms, each registered to a different sub-volume of the problem space. The implemented algorithms include Patch-based Multi-View Stereo, Example-based Photometric Stereo, Gray-code Structured Light, and Volumetric Visual Hull;
• Developed an algorithm & evaluation pipeline that runs algorithms on both real-world and synthetic datasets, compute and visualize quantitative results for evaluation.

• Implement and maintain an open source Multiple View Geometry library with minimum third-party dependencies;
• Implement incremental Structure from Motion (SfM) using open3DCV.

• Implement sequential Structure from Motion (SfM) using open3DCV and write a detailed tutorial for educational purpose.

• Developed 3D reconstruction algorithms and high-level abstraction for everyday objects;
• Multi-projector calibration, header tracking using kinect and magnetic tracker.

• Implemented from scratch a robust cylindrical image stitching algorigthm in C++, which includes feature detection/extraction, focal length estimation, planar plane to cylindrical surface transformation, feature-based image alignment algorithms, etc.

• Course project of CPSC 554M - Topics in Human Computer Interaction. Personalize video thumbnails based on viewing behaviour for fast video re-searching.

• Course project of EECE 418 - Human Computer Interface. A travel agent APP that uses online social connections and common interests to match potential backpackers.

• Use decision stumps and decision trees as base learners, implemented different split criteria for the decision tree, including information gain of ID3 and information gain ratio of C4.5;
• Implement post-prune for the tree construction algorithm using reduced error pruning;
• Implement one-vs-rest multi-class classification using decision stumps as base learners, and SAMME using decision trees as base learners.

• Investigate the performance of Adaboost when three different weak classifiers - decision stump, decision tree, and weakened SVM are used as the base classifier.

• Implemented several pixel-based and color histogram-based shot boundary detection algorithms with an GUI interface for easy browsing.

• Hand gesture segmentation using color and depth cues, and recognition using extracted hand geometric features.

• Implemented independently an iClicker-like Android APP to send/receive information, and a PC program that acts as a server using multi-thread Socket and has information processing functionalities.

Instructor: Jim Little, Ph.D.

Held weekly tutorial, assignments and exams marking

Instructor: Fredrick Tung, Ph.D. candidate

Held weekly tutorial, assignments and exams marking

Instructor: Konstantin Beznosov, Ph.D.

Manage two lab sessions, help students with coding projects, held weekly scrum meetings, assignments and exams marking

Second-class scholarship

Recipient of the Second Prize (Top 5%)

Won runner-up of the Most Popular Exhibit award

Recipient of Successful Participant

C/C++ • MATLAB • Java • Python

OpenCV • OpenGL • Eigen • Ceres Solver • openMVG • TheiaSfM • libmv • PCL • LaTeX • Blender

Guitar, Photography, Music, Movies, Reading, Cooking.