Differences

This shows you the differences between two versions of the page.

--- projects [2011/09/02 21:47] – dymond
+++ projects [2013/04/29 12:27] – mb
@@ Line 1: / Line 1: @@
-====== Currently offered Projects, Fall 2011 (updated September 2, 2011)  ======
+====== Available Projects  ======
-(Listed in order received.)
-====== Building an autonomous motorboat ======
+Currently offered Projects, Summer 2013
-**Supervisor**: Michael Jenkin
+(some projects still subject to confirmation)
-**Required Background**: General CSE408x prerequisites
+======Tracking and Activity Recognition Through Consensus in Distributed Camera Networks======
-**Recommended Background**: Robotics
+**Supervisor**: Amir Asif
-__Description__
+**Required Background: ** Computer Vision or Signal and Systems Course preferred; Matlab; Interest in Signal/Image Processing
-An opportunity exists for a small number of students to build an autonomous motorboat using a RC motorboat as a base and integrating computation and control in the form of a Beagleboard. Students will participate in lectures and labs associated with CSE6324 (Part I). Interested students should attend the first lecture of CSE6324. See the departmental schedule for time and place.
+**Short Description: **Over the past decade, large-scale camera networks have become increasingly popular in a wide range of applications, including: (i) Sports analysis; (ii) Security and surveillance; (iii) disaster response, and; (iv) Environmental modeling, where the objective is to follow the trajectory of a key target, for example, a star player in a soccer game or a suspect in a surveillance environment. In many applications, bandwidth constraints, security concerns, and difficulty in storing and analyzing large amounts of image data centrally at a single location necessitate the development of distributed camera network architectures. In this project, we investigate distributed scene analysis algorithms, where each camera estimates certain parameters of the target using a signal processing algorithm based upon its own set of observations. The local estimates are then shared with the neighboring cameras in an iterative, goosip-type fashion, and a final estimate is computed across the network using consensus algorithms. The selected student will develop Matlab code to apply distributed signal processing algorithms [1,2] that have been developed in the Signal Processing and Communications lab for target tracking and activity recognition in distributed camera networks.
-----
-:
-----
-====== Athenians Data Project ======
+[1] A. Mohammadi and A. Asif, Distributed Particle Filter Implementation with Intermittent/Irregular Consensus Convergence, IEEE Transactions on Signal Processing, 2013. http://arxiv.org/abs/1112.2431.
-**Supervisor**: Nick Cercone
+[2] A. Mohammadi and A. Asif, Decentralized Sensor Selection based on the Distributed Posterior Cramer-Rao Lower Bound, in proceedings of IEEE International Conference on Information Fusion, Singapore, 2012. pp. 1668-1675.
-**Required Background**: General CSE408x prerequisites
+====== 3D Drawing System with Leap Motion finger tracker ======
-**Recommended Background**: Data Mining
+**Supervisor**: Wolfgang Stuerzlinger
-__Description__
+**Required Background**: 3D computer graphics, C/C++ coding
-The Athenians Project is a multi-year, ongoing project of compiling, computerizing and studying data about the persons of ancient Athens.
-Possible project ideas for this term span from simpler ones such as
-how to present data in the best possible way, add spatial characteristics to existing data,
-add multimedia data, improve text searching, etc. to more complex ideas such as filling
-missing parts for the "broken" words on the existing inscriptions. Filling text for the broken
-words has been done in the past using expert knowledge. Those experts have establish
-certain rules/guidelines that may be possible to extrapolate in some kind of expert system
-when talking in IT terminology. Furthermore, any hypotheses on word completion enters
-the database with some likelihood. Associating probabilities with hypotheses introduces
-another opportunity for research projects.
-----
-:
-----
-====== Three-Dimensional Context from Linear Perspective for Video Surveillance Systems ======
+The Leap Motion, leapmotion.com, is a new device that lets users control a computer with their fingers. This project creates a new 3D drawing system that enables users to quickly generate 3D solids.
-**Supervisor**:  James Elder
+====== 3D Drawing System with 3Gear gesture tracker ======
-**Requirements**:  Good facility with applied mathematics
+**Supervisor**: Wolfgang Stuerzlinger
-__Description__
+**Required Background**: 3D computer graphics, C/C++ coding
-To provide visual surveillance over a large environment, many surveillance cameras are typically deployed at widely dispersed locations.  Making sense of activities within the monitored space requires security personnel to map multiple events observed on two-dimensional security monitors to the three-dimensional scene under surveillance.  The cognitive load entailed rises quickly as the number of cameras, complexity of the scene and amount of traffic increases.
+The 3Gear system, threegear.com, lets users control a computer with their hands and fingers. This project creates a new 3D drawing system that enables users to quickly generate and modify 3D solids.
-This problem can be addressed by automatically pre-mapping two-dimensional surveillance video data into three-dimensional coordinates.  Rendering the data directly in three dimensions can potentially lighten the cognitive load of security personnel and make human activities more immediately interpretable.
-Mapping surveillance video to three-dimensional coordinates requires construction of a virtual model of the three-dimensional scene.  Such a model could be obtained by survey (e.g., using LIDAR), but the cost and time required for each site would severely limit deployment.  Wide-baseline uncalibrated stereo methods are developing and have potential utility, but require careful sensor placement, and the difficulty of the correspondence problem limits reliability.
-This project will investigate a monocular method for inferring three-dimensional context for video surveillance.  The method will make use of the fact that most urban scenes obey the so-called “Manhattan-world” assumption, viz., a large proportion of the major surfaces in the scene are rectangles aligned with a three-dimensional Cartesian grid (Coughlan & Yuille, 2003).  This regularity provides strong linear perspective cues that can potentially be used to automatically infer three-dimensional models of the major surfaces in the scene (up to a scale factor).  These models can then be used to construct a virtual environment in which to render models of human activities in the scene.
+====== Tilt Target Selection on Touchscreen Phones ======
-Although the Manhattan world assumption provides powerful constraints, there are many technical challenges that must be overcome before a working prototype can be demonstrated.  The prototype requires six stages of processing:    1)The major lines in each video frame are detected.  2)  These lines are grouped into quadrilaterals projecting from the major surface rectangles of the scene.  3)  The geometry of linear perspective and the Manhattan world constraint are exploited to estimate the three-dimensional attitude of the rectangles from which these quadrilaterals project.  4)  Trihedral junctions are used to infer three-dimensional surface contact and ordinal depth relationships between these surfaces.  5)  The estimated surfaces are rendered in three-dimensions.  6)  Human activities are tracked and rendered within this virtual three-dimensional world.
+**Supervisor**: Scott MacKenzie
-The student will work closely with graduate students and postdoctoral fellows at York University, as well as researchers at other institutions involved in the project.  The student will develop skills in using MATLAB, a very useful mathematical programming environment, and develop an understanding of basic topics in image processing and vision.
+**Required Background**: General 4080 prerequisites, CSE3461, and (preferably) CSE4441. Interest in user interfaces and human-computer interaction (HCI).  Students can use their own Android phone for the project or one supplied by the course supervisor.
-For more information on the laboratory: [[http://www.elderlab.yorku.ca]]
+Touchscreen mobile devices commonly use a built-in accelerometer to sense movement or tilting actions of the device.  Tilt is commonly used the change the orientation of the display between portrait and landscape.  Gaming is another common use for tilting actions.  However, tilt may also be used for target selection, as a replacement for touch.  This research project will evaluate tilt as an input primitive for target selection on touchscreen mobile devices.
-----
+**Readings**: MacKenzie, I. S., & Teather, R. J. (2012). FittsTilt: The application of Fitts’ law to tilt-based interaction. Proceedings of the Seventh Nordic Conference on Human-Computer Interaction – NordiCHI 2012, pp. 568-577. New York: ACM.
-:
-----
-====== Estimating Pedestrian and Vehicle Flows from Surveillance Video ======
+====== Attentive Sensing for Better Two-Way Communication in Remote Learning Environments ======
-**Supervisor**:  James Elder
+**Supervisor**: James Elder
-**Requirements**:  Good facility with applied mathematics
+**Required Background**: General CSE408x prerequisites, good programming skills,
+good math skills, knowledge of C and MATLAB programming languages
-__Description__
+One of the challenges in remote learning is to allow students to communicate effectively with the lecturer.  For example, when a student asks a question, communication will be more effective if the instructor has a zoomed view of the student’s face, so that s/he can interpret expressions etc.
+The goal of this project is to apply attentive sensing technology (www.elderlab.yorku.ca) to this problem.  This technology is able to monitor a large environment such as a classroom and direct a high-resolution ‘attentive’ sensor to events of interest.
+In particular, working with a senior graduate student or postdoctoral fellow, the  successful applicant will:
+  - Study the problem of detecting hand-raises in the preattentive sensor stream
+  - Implement algorithms for detecting hand-raises based upon this investigation
+  - Evaluate these algorithms in a real-classroom setting, using proprietary attentive sensing technology
-Facilities planning at both city (e.g., Toronto) and institutional (e.g., York University) scales requires accurate data on the flow of people and vehicles throughout the environment.  Acquiring these data can require the costly deployment of specialized equipment and people, and this effort must be renewed at regular intervals for the data to be relevant.
-The density of permanent urban video surveillance camera installations has increased dramatically over the last several years.  These systems provide a potential source of low-cost data from which flows can be estimated for planning purposes.
+====== Attentive Sensing for Sport Video Recording Markets ======
-This project will explore the use of computer vision algorithms for the automatic estimation of pedestrian and vehicle flows from video surveillance data.  The ultimate goal is to provide planners with accurate, continuous, up-to-date information on facility usage to help guide planning.
+**Supervisor**: James Elder
-The student will work closely with graduate students and postdoctoral fellows at York University, as well as researchers at other institutions involved in the project.  The student will develop skills in using MATLAB, a very useful mathematical programming environment, and develop an understanding of basic topics in image processing and vision.
+**Required Background**: Good programming skills; Good math skills; Knowledge of C and MATLAB programming languages
-For more information on the laboratory: [[http://www.elderlab.yorku.ca]]
-----
+The goal of this project is to modify York University’s patented attentive sensor technology to the sport video recording market.  Specific application domains under investigation include skiing, indoor BMX parks, and horse tracks.
-:
-----
+The general problem is to use attentive sensing technology (www.elderlab.yorku.ca) to visually detect and track multiple moving agents (e.g., skiers, riders, horses) and to select specific agents for active high-resolution smooth pursuit.
+The student will work with senior graduate students, postdoctoral fellows and research scientists to help modify the attentive sensing technology to operate in these domains.   Specific tasks include:
+.     Ground-truth available datasets
+.     Evaluate current attentive algorithms on these datasets
+.     Modify these algorithms to improve performance on these datasets
-====== Tandem repeat detection using spectral methods ======
-**Supervisor**: Suprakash Datta
+====== Continuation of a Path Diagram to Syntax Application ======
-**Required Background**: The student should have completed undergraduate courses in Algorithms and Signals and Systems.
+**Supervisor**: Jeff Edmonds
-**Recommended Background**: Some background in Statistics is desirable but not essential.
+**Required Background**: General CSE408x prerequisites
-__Description__
+**Recommended Background**: Java software development
-DNA sequences of organisms have many repeated substrings. These are called repeats in Biology, and include both exact as well as approximate repeats. Repeats are of two main types: interspersed repeats (which are spread across a genome) and tandem repeats, which occur next to each other. Tandem repeats play important roles in gene regulation and are also used as markers that have several important uses, including human identity testing.
-Finding tandem repeats is an important problem in Computational Biology. The techniques that have been proposed for it fall into two classes: string matching algorithms and signal processing techniques. In this project, we will explore fast, accurate algorithms for detecting tandem repeats and evaluate the outputs of the algorithms studied by comparing their outputs with those of available packages, including mreps (http://bioinfo.lifl.fr/mreps/), SRF (http://www.imtech.res.in/raghava/srf/) and TRF (http://tandem.bu.edu/trf/trf.html).
+Structural equation modeling (SEM) is a statistical technique that is becoming increasingly popular in the educational and behavioral sciences. SEM allows researchers to test the validity of hypothesized models involving complex relationships among multiple variables. Collected data is used to estimate the parameters of the equations and assessing the fit of the model.
-The student will implement existing spectral algorithms based on Fourier Transforms and on an autoregressive model. He will then make changes suggested by the supervisor, and evaluate the effect of the modifications. Throughout the course, the student is required to maintain a course Web site to report any progress and details about the project.
+The software required is an application that allows researchers to define their hypothesized models visually and will output the correct syntax for the analytical software of their choosing.
+To date a promising functional application has been developed in JAVA by a Computer Science student as a 4080 project. The existing software allows the user to draw a path diagram and outputs code for the R package sem.  There are a number of improvements to be made (refinements and additions to graphical user interface) and then the application needs to be extended to output syntax appropriate for additional software applications (openMX, MPlus and EQS).  Though this project may not begin at “the first stages” of the software lifecycle, this scenario is likely common in the software development market. In addition, the student will be working with a primary “client” who is far less technically advanced, which is also reflective of real-world situations.
+More details {{:continuation_of_a_path_diagram_to_syntax_application.pdf|here}}.
-----
-:
-----
-====== Touch- and Gesture-based Text Entry With Automatic Error Correction ======
+====== Enabling SaaS access to an experimental AI planner ======
-**Supervisor**: Scott Mackenzie
+**Supervisor**: Sotirios Liaskos (liaskos at yorku dot ca)
-**Required Background**:
+**Required Background**: Good knowledge of Unix tools / Python, Perl or Awk. Comfort with algorithms and programming. Essential: 2031 -- Software Tools. Desired:
-CSE3461 (or equivalent), CSE3311 (or equivalent), CSE4441 (or equivalent)
+-- Functional & Logic Programming,
-A student wishing to do this project must be well versed in Java, Eclipse, and developing java code for the Android operating system.
+-- Design and Analysis of Algorithms,
+-- Compilers and Interpreters.
+**Description**: This project involves enriching and integrating a set of fairly complex scripts, which are components of an Artificial Intelligence (AI) planner, and exporting them to the public in a Software-as-a-Service (SaaS) fashion.
-**Recommended Background**:
+The components are various Unix executables and LISP programs that need to interact in complex ways. The components may be residing in different servers in different universities. Currently integration is performed manually, at the expense of usability. Thus, we aim at constructing a module that: (a) integrates involved components to deliver output in one call, (b) exports a unique web interface (preferably following WSDL/SOAP) to be easily accessed by custom front-end tools by anyone, anywhere, (c ) offers a simple front-end for human users.
-Possession of an Android touch-based phone or tablet would be an asset, but is not essential.
-__Description__
+Learning objectives:
-This project involves extending a touch-based text entry method to include automatic error correction.  The method, as is, uses Graffiti strokes entered via a finger on a touch-based Android tablet.  The stroke recognizer works fine, but it is not perfect.  Some strokes are mis-recognized while others are un-recognized.  The fault is sometimes attributable to the recognizer, but, often, the fault is simply that the user's input was sloppy.  The work involves developing, integrating, and testing software.  The core software is already written, but automatic error correction is lacking. The primary task of the added software is to receive a sequence of characters representing a word and matching the sequence with words in a dictionary.  If a match is found, all is well (presumably).  If a match is not found, the search is extended to find a set of candidate words that are "close" to the inputted sequence.  "Close", here, involves using a minimum string distance algorithm (provided).  The user interface must be modified to present the user with alternative words in the event an error occurred.  The user selects the desired word by tapping on a word in the list.  The project will involve testing the new input method in a small user study and writing up a report describing the work and presenting the results of the user study.
+  * Understand the technologies and process involved in turning native code into a web-service ("servicizing").
+  * Study a state-of-the-art AI planner and understand its workings.
+  * Exercise scripting skills.
+====== Predicting Angular Error in Rigid Registration ======
+**Supervisor**: Burton Ma
+**Description**: Registration is a fundamental step in image-based surgical
+navigation. Several (seemingly) different approaches for predicting
+distance errors in registration are known, but for some surgical
+procedures, the angular error in registration is more important.
+This project will validate an approach for predicting angular
+error in registration; the student will use a combination of
+simulated and actual registration data for testing purposes.
+====== Calibration of a Tracked Pointer ======
+**Supervisor**: Burton Ma
+**Description**: Tracked pointers are the most common tools used in surgical
+navigation systems. A typical pointer has a tracked target on one
+end and a sharp or ball tip on the other end. Finding the location
+of the tip relative to the target is a calibration problem. One
+solution to the calibration problem involves pivoting the pointer
+about the tip while tracking the target; if the tip is kept
+stationary, then the target moves on the surface of a sphere.
+Fitting the tracking data to the surface of a sphere yields the
+location of the tip as the sphere center. Unfortunately, the
+calibrated tip position obtained using such a spherical calibration
+has high variance. This project will investigate how much variance
+there is in the calibrated tip position, and methods for reducing
+the variance of the calibrated tip position.
+====== A privacy safeguard framework for sharing photos on Facebook ======
+**Supervisor**: Uyen Trang Nguyen
-----
+**Description**:
-:
+One of the major privacy concerns in Online Social Networks is photo sharing.  A user may post his/her friends’ photos without their consent.  The friends have no control over the user’s Facebook activities, namely photo sharing.  In this project, we design and implement a third-party Facebook application that allows people to protect their identities in photos uploaded by another user without their consent.
-----
+**Required prerequisite background**:  Proficiency in programming, especially in Java and Web application programming.
+**Desired prerequisite**: Knowledge of image processing, Facebook API, JavaScript Object Notation (JSON)