Upcoming Courses • Fall 2016: “Exploratory Research” • Spring 2017: “Advanced Topics in Computer Vision” Exploratory Research Sign up for EENG/CSCI 599 (“Independent Study”) with me as the advisor • Purpose: Exploratory research on computer vision‐ related projects from industry or in an area of mutual interest • Similar to our final projects, but semester‐long • Benefits to students: • Learn how to develop ideas, explore solutions, develop prototypes, do experiments • Interact with potential employers; gain expertise in a topic area of interest to them • Potential for commercialization; student owns any intellectual property (IP) developed Possible Project – nVoq, Inc • Computer Vision Enabled Desktop Workflow Automation • Goal: Locate components such as text fields, buttons, boxes on the computer screen. Read the text in the text field. • Motivation: Desktop workflow automation can simplify and speed up lengthy repetitive tasks that often cross application boundaries by scripting the operations. For applications that don’t expose their functionality via an API, we need computer vision. Possible Project – Apex Mtn Eng • License Plate Recognition • Goal: Locate and read license plates from a dashboard mounted camera and an embedded processor such as Raspberry pi • Motivation: Law enforcement can benefit from a system that detects license plates corresponding to stolen cars, amber alert vehicles, etc Possible Project – Healing Waters • Analog Gauge Reading • Goal: Automatically recognize analog gauges of water pumps and determine their settings • Motivation: Company has a large number of units installed in poor areas of Africa, and needs a way for locals to collect data on the pumps. Ultimate goal is to port algorithm to a cell phone. Advanced Topics in Computer Vision (Spring 2017) • EENG/CSCI 508 • Builds on the course you are taking now • Topics: • • • • • RGB‐D (i.e., 3D) cameras and image processing Activity recognition Stereo and dense reconstruction from multiple views Multiview motion and structure estimation (SLAM) Advanced object recognition, including category recognition • Advanced segmentation methods