User:Reemabajwa

Through my life experiences, I have become a firm believer in the effect a person’s name has on their personality. My father named me Reema, a Persian word used for a gazelle. Although I may not be able to compete with the elegance and speed of this antelope specie, I do feel that the death of my father when I was six years old, followed by a life as part of an all women family dwelling in a patriarchal society, has trained me to be the gazelle I was named after- to be elegant in forming solutions and quick in solving them. It is this acquired urge of breaking down every problem into solvable modules that binds me to Computer Science and its inherently similar nature. In this spirit of solving problems, I wish to develop interaction centric technology that bridges the gap between humans and computers. A graduate degree in the MS CS program at the Cornell University would allow me to fulfill this passion by augmenting my skill-set and developing me professionally. As a computer science undergraduate student at Lahore University of Management Sciences (LUMS), I discovered a passion for Human Computer Interaction by exploring the avenues of research and practical applications. In my freshmen year, under the supervision of Dr. Sohaib Khan, I carried out research projects on Microsoft Kinect and developed a program that allowed control of presentations through human gestures. In my sophomore year, I developed an RMI based interactive game that measured the similarity between real time drawings. The logic of the game resided in edge and shape detection of objects within images which were extracted by implementing the canny edge detector and compared using normalized compression distance. Shortly afterwards, I developed a 3-D bowling game integrated with a Wii remote that aimed to provide users with a real life bowling experience. The intensity of the bowling motion, captured through the acceleration of the Wii mote, was used to render graphics accordingly. After these experiences of designing seamless interactive mechanisms and detailed user experiences between humans and computers, I felt hI had explored an avenue that promised me the opportunity to channel the creativity that I had always displayed in my sketches, acrylic paintings as well as graphic designing skills.

While at LUMS, the passion for interactive technology drove me to explore the domains of Computer Graphics, Computer Vision and Digital Image Processing. In my Computer Graphics course, I successfully developed a Ray Tracing and Photon Mapping image renderer that rendered a given scene according to the lights and object materials present within. In fall 2014, I worked a teaching assistant for the same course, and designed assignments on open GL, conducted sessions on Unity3D, and overlooked research projects on hair and particle simulations and fractal graphics. Furthermore, as a course project for Computer Vision, I researched on and developed an application that reconstructed a 3-D environment (with Manhattan world assumptions) captured by a single photograph. The camera calibrations were obtained by extracting the focal length from computations on vanishing lines and the planes were reconstructed using parallel lines marked by a user through an interactive GUI. In the domain of Digital Image Processing, I researched on Content and Semantics Based Image and Video retrieval. In this project, I combined multiple approaches such as Gaussian fitting and computation of binary feature vectors based on entropy changes to devise an effective approach of video retrievals. I then extended this approach to research on an optimal balance between efficiency and accuracy in distributed systems (especially peer to peer systems) by combining a super peer network with binary spatiotemporal feature vectors. In the summer of 2014, I worked at the Computer Vision Laboratory through a research-based internship with Dr. Murtaza Taj. Essentially, I improved the dome fitting algorithm by using b-splines instead of the formerly used polynomial function, and by using surface revolution to transform a profile curve into a 3-D primitive. Digging deeper into the world of point clouds, for my senior year project, I chose to work on the reconstruction of 3-D CAD like models from noisy point clouds obtained through laser scanners under the guidance of Dr. Sohaib Khan and Dr. Murtaza Taj. We extended previous work on segmentation of point clouds and developed a novel n-gon algorithm designed to fit 3-D n-gonal polyhedrons of multiple radii (where n is the number of sides of a given primitive) to noisy point clouds that are first projected in a 2-D space allowing for greater efficiency and accuracy of fitting. The number of sides of a regular polygon was automatically obtained by clustering point cloud normals and extracting prominent segments. After this work was published in EuroGraphics 2015, we secured a hefty funding from USAID, through which we purchased a Leica P20 Laser scanner, thus eliminating the issues of obtaining real local data for the first time in Pakistan! Four months down the project, Dr. Taj and I have set up an entire team, which is actively working towards the goal of scanning and modeling six major heritage sites in Pakistan. Since I graduated in June 2015, I have been working with Dr. Taj on devising approaches to extract fine details from point clouds instead of the bare minimum parameters that fit coarse primitives on a segmented region of the point cloud. By first processing the segmented point cloud data to a one dimensional signal (either profile curves, or 2-D projections on the ground plane) I have used peak finding and clustering techniques to extract parameters that define the structural changes in a primitive. These parameters are then given as the only required input to a personalized library of heritage architectural primitives, which is designed to generate a given primitive using its respective mathematical equations. Moreover, I have been guiding two junior students on a research project based on global fitting of an automatically generated 3-D model consisting of several primitives. We are currently writing an optimization that takes into account several constraints including orthogonality, parallelism and symmetry while ensuring proximity to the real point cloud.

Apart from my academics and research, I have also been involved in entrepreneurship in order to convey my technical skills, designing abilities and pro-activeness to the community around me in a professional manner. In September 2013, as part of a team of six members, I won the first prize at Pakathon, a Boston based startup hackathon arranged by the MIT enterprise board. With the seed money obtained from the competition, my team and I set out to develop ‘lookout360’, an interactive map based interface that allows users to report a range of events as well as view a wider picture of activities in their communities in a visually simplistic manner. Furthermore, I have also co-launched a software development company by the name of Walzip whereby I design and develop interactive websites and visualization applications. From the time it was established in June 2014, Walzip has successfully developed a number of products for both local and international customers. In August 2014, Walzip offered its first internship program for O and A’ Level students who were required to prepare and organize content for a GCE O’Level academic application. My career goals are to continue researching in the area of HCI, Computer Vision and Graphics and to amalgamate my findings into applications/devices that have an impact on the community at large. Being a Google Anita Borg Scholar and a Google Student Ambassador, I have initiated a campaign for the empowerment of women in technical professions. Moreover, I have also enjoyed working as a Teaching Assistant for Computational Problem Solving and Computer Graphics at LUMS. These endeavors have made me realize that spreading the knowledge of Computer Science, especially to the women of my country, is as important of a goal to me as is gaining the very same knowledge and is something that I wish to take forward at every step of life.