Technical lead for the SmartSight project. Lead a team of 12 testers validating new features on the product including navigation and computer vision algorithms. Team successfully completed 3 long term POVs, deploying and stabilizing the system at customer sites.

Provide various technical services including camera system design, electronic design, embedded firmware, algorithm development.

-One of three initial employee.-Main responsibility is firmware that controls the image/video capture pipeline.Improved image quality throughout our workflow - from on the camera to cloud-based 
functions.-Led entire hardware team consisting of 1-2 direct reports.-Led hardware verification and testing on two versions of the camera.-Provided support to manufacturing including instructions, specifications and troubleshooting.-Worked with outside manufacturer to design the lens focus station and end-of-line tester used in camera production.-Worked with mobile app and web development team to ensure hardware functionality could be accessed by various services.-Advised and consulted with image processing research team; initiated colour balancing and noise removal process for cloud deployment.-Involved in hardware design of new version of camera including component selection and architecture.

-2 month project-Main objective is to explore the potential of adding quantitative measurement of joint kinematics to the Mobility Assessment System (MAS) by incorporation of a Microsoft Kinect sensor into the system. The Kinect’s skeleton detection algorithms offer an easy way to detect a remarkable number of body postures via 20 joints.-The goal of the project is the automation of the walking and standing portions of the Tinetti test.-Responsible for translating subjective criteria of the Tinetti Assessment Test into a set of quantitative measurements based on the skeleton extraction of the patient from a Kinect sensor.

-Course: ENG 3000 - Professional Engineering Practice-Present course material to 3rd year engineering students-Responsible for creating slides, assignments and exams.

Project Overview:The purpose of the ARTSWARM project is to study the behaviour of collective animal groups. In particular, we are interested in the collective movement of starling flocks and midge swarms. The motivation for these displays is quite different; starlings flock as a defensive mechanism against predation, while midge swarms are a mating ritual for male midges to attract female midges. The understanding of the fundamental dynamics of these biological events requires more than simple observation and reporting. For this project, empirical data in the form of stereometric photographs have been captured, which form the basis for numerical analysis. The main challenge in this research is that these biological displays cannot be easily replicated in a laboratory environment and can only be obtained in the field. Hence, the uncertainties of the environment (weather, movement of the sun, disappearance of animals due to migration or death) and the probability of capturing images of a relevant event are the limiting factors, rather than more technological issues (image segmentation, target tracking).Contributions:-Part of experimental team responsible for capturing high speed stereometric photographs of collective animal behaviour (starling flocks and midge swarms).-Developed intrinsic and extrinsic camera calibration procedure and associated software required for full calibration of stereo camera setup making metrology possible.-Responsible for verification testing of stereo camera system to ensure proper reconstruction of captured collective behaviour events.-Developed software to automatically verify the validity of 3D trajectories, thus providing instant feedback on the quality of the tracking algorithm.-Verification testing of image segmentation and tracking algorithms, which lead to improvements to both methodologies.

Research Overview:My research was focused on optimization of resources in distributed smart camera (DSC) networks. In particular, a method to allow individual nodes/cameras be responsible for decision-making about their own field-of-view (FOV). Inspiration for the new method was drawn from distributed systems in nature such as ant colonies and bird flocks. In these systems, individuals follow simple rules, without centralized control, while trying to complete a certain task. The result of each of the individuals' actions is an emergent system-level behaviour, which is governed by the principles of self-organization. Applying these principles to DSC networks, a camera selection methodology that requires no centralized control has been designed. Using only the local quality-of-view (QOV) metric provided by each camera and communication with a small number of other cameras, the camera selection methodology has been shown to optimize a set of system-level metrics. The calculations required by the methodology are not processor intensive and communication between cameras is limited.Contributions:-Designed and built twelve camera inward-looking distributed smart camera network using low-cost and low-power components.-Created and validated a novel decentralized camera selection methodology based on the principles of self-organization, which requires very little network bandwidth.Created a novel camera calibration target and implemented corresponding procedure that requires only a single image per camera versus 20+ images for a typical method.-Supervised multiple undergraduate summer students, whose work contributed directly to several publications.-Involved in many other projects such as: scale factor camera calibration; CMOS image sensor system-on-chip design; high dynamic range pixel design; temperature effects in CMOS image sensors; halo measurements in night vision goggles; and triangulation rangefinding.

-Performed “1st silicon” (first production) verification and validation tests on several new CMOS camera products within 1 week, hence ensuring first-to-market position. Wafer-level tests included: image quality (noise, colour, blooming), subsystem testing (analog and digital gain, onboard memory) and temperature working range validation-Created automated data analysis and report generation code in MATLAB for presentation of “1st silicon” test results. This package processed raw test data and produced not only graphs and figures-of-merit, but also automatically generated powerpoint slides for the final report. The result was reducing personnel time spent on creating and formatting the presentation, leading to a reduction in the overall time required to complete 1st silicon verification.-Learned to operate probe stations, micro probes and other specialized test equipment, which allowed me to work independently with little or no supervision.-Investigated image quality issues such as MTF, dead pixel clusters and noise, leading to better analytical tools to quantify these yield issues.

-Surveyed communication sites and drafted as-built technical diagrams providing up-to-date documentation for future expansion and/or customization.-Communicated with local engineers and maintenance technicians about improvements that could be made to each site diagram.

-Researched and authored a feasibility study on the installation of a microwave radio link, including microwave radio path design, equipment lists and cost analysis.-Researched, designed and programmed a PLC system for monitoring vital conditions (humidity, temperature, battery condition) of unmanned communication sites, reducing the number of maintenance trips required.-Completed VHF radio path designs for ATCO Pipelines, leading to more robust communication channels.Surveyed communication sites and drafted as-built technical diagrams.-Organized and coordinated the installation of satellite phones and radio remotes as part of Y2K preparations ensuring the availability of emergency communications.

-Performed integration radio frequency radiation hazard (RADHAZ) tests on entire fleet of IRIS System-equipped Canadian military vehicles.-Performed design proving tests on antenna configurations and RF data transmission procedures.-Drafted full RADHAZ test result reports for integration and qualification level testing, which made recommendations concerning RADHAZ dangers to the Canadian Government.