The annual Computer Science & Engineering
Career Fair and Poster Competition was held at Texas A&M University on September 8, 2015. Members of the Sketch Recognition Lab - both
graduate and undergraduate - presented their projects to experts and to future employers ranging from American Airlines to Facebook.
(Director of SRL, Dr. Tracy Hammond [left] and Stephanie Valentine [right] at the CSE Career Fair)
([Left to right] Swarna Keshavabhotla, Jung-In Koh, Seth Polsley, Anurag Garg, and Trevor Nelligan at the CSE Fair)
Among the posters of CSE students were many presenting SRL projects and student research.Raniero Lara-Garduno showcased a poster about updated information regarding Smart Strokes, his project that allowed for the analysis and evaluation of digitalized neuropsychology tests. Normally done with pen and paper, Lara-Garduno aims to modernize the method used in these connect-the-dot tests to take advantage of sketch recognition tools and technology, which allows for greater understanding to come from the tests.
Josh Cherian presented his research on action recognition during cleanliness activities (for example, brushing teeth or washing hands). Currently, Cherian is focused on applying his findings and project to those in the elder community, who may have cognitive decline due to aging or illness. Using an Andriod app and a Pebble Smartwatch, Cherian was able to track and differentiate different movements and match them to cleanliness activities. This research can help aides keep track of patients who need help remembering whether or not they washed their hands or brushed their teeth.
(Josh Cherian at the IAP Poster Competition)
Siddartha Karthik's presentation was in regards to picture and sketch labeling. His goal is to create an effective and simple way to label, store, and retrieve drawings entered into programs via pen-and-touch technology. Karthik and SRL have already created an automatic drawing labeling program. It allows users to enter sketches into the program and receive a list of most probable labels based on what is recognized. Ultimately, this program would allow for more efficient organization and recall ability for hand drawings.
Also at the contest, Ayden Kim reported his work on classifying and assessing motor skills in children to predict whether they were ready for school. Previous research has shown that fine motor control is related to school and study readiness, and Kim aims to revamp the current motor 'Star Test' into a digital and touch interface, KimCHI. By drawing various, simple shapes, KimCHI evaluates whether a child's motor skills are 'mature' and ready for school and studies, or 'in training' and needing more work.
Presenting his new project at the poster competition was Vijay Rajanna. His research involves eye tracking during computer-human interaction (CHI). Using a gaze tracking device and a pressure pad-like device he engineered, Rajanna is working on creating a new way of using computers without the use of a traditional computer mouse. His new method of CHI would allow for more efficient computing and precise interaction with very little practice. While his work is still in the early stages of research and testing, Rajanna hopes that his future work will make CHI more interactive and less limited.
Swarna Keshavabhotla presented a poster on Persketchtivity. This project is a web-based application that aids in practice drawing geometric shapes for engineering students. Persketchtivity helps students by offering users outlines to trace using pen-and-touch technology. Once the user traces the shape, Persketchtivity shows the differences between the inputted sketch and the original outline. At the moment, Keshavabhotla is working on adding more advanced shapes and finessing the program.
(Keshavabhotla presenting at the IAP Poster Competition)
Purnendu Kaul presented his new research regarding Mechanix. Kaul is currently combining gaze-tracking technology and the intelligent tutoring system created in SRL (Mechanix) to better understand user's state of mind while solving problems. In Mechanix, students can practice drawing and solving truss problems and gain feedback on their work. By tracking their gaze while working, Kaul can see where a user's gaze lingers and/or returns to (for example, when a person reads part of a problem more than once or concentrates on just one word).
Jung In Koh showcased her work on gesture recognition and emoticons in regard to online expressiveness. Through interviews and online surveys, Koh is researching trends in intuitive gestures to describe common emojis. By recording interviews and using gesture tracking technology, Koh has found that there are five emoticons that coincide with a consistent gesture - for example, the thumbs up icon. Moving forward, Koh is interested in continuing her project to experiment in algorithms that allow for emojis to be connected to gestures.
Jaideep Ray's poster presentation was about his work in sketch matching, recognition, and retrieval. Ray's research is centered around the goal of more efficient image tagging - that is, a more efficient way to digitally classify and identify what a drawing represents. Instead of being identified by where lines intersect, Ray is working on applying an algorithm that focuses on sketch recognition through spatial markers and distance between lines. This would allow for more effective recall of images and objects in sketch recognition programs.
Paul Taele competed with a poster about his program, BopoNoto. As a tool for learning how to write languages that use zhuyin phonetic script, BopoNoto prompts users with an English spelling of a phonetic sound and recognizes the sketch that is input into the application. BopoNoto not only corrects the overall sketch, but also the stroke order, which is important to many non-Romanized languages. Taele's program allows users to learn how to correctly write in different languages, and his future work looks to better the recognition system as well as expand his program to other world languages.
David Turner and Matt Runyon worked in tandem to present a poster on CourseSketch. Their ongoing electronic gradebook is aimed to make a more intuitive, effective system for educators to use. CourseSketch would help educators program more advanced policies into the gradebook - such as late policies - that would automatically be applied to grades, help educators see grade trends, and aid students in determining how they are doing in their classes. Going forward, Turner and Runyon want to add innovative features to the system, such as graphic visualization of grade statistics, and use artificial intelligence to make a more advanced system.
(David Turner and Matt Runyon presenting at the CSE Career Fair)
SRL and Director Dr. Tracy Hammond are also excited to congratulate Trevor Nelligan for winning first place in the poster competition!!
(Left: Nelligan presenting his poster. Right: Dr. Hammond and Nelligan after the first place announcement)
Trevor presented a poster on Mechanix for the IAP contest. He talked enthusiastically to the judges about the educational interface. Mechanix, which is a sketch-based tutoring system for engineering students enrolled in introductory statics courses, allows for professors to give drawing assignments to hundreds of students at a time. The system gives question scenarios, recognizes geometric object drawings and force arrows, and gives instant feedback to users. This allows for students to both practice engaging assignments while also explaining any errors in calculations that happened along the way.
We are proud of everyone who participated in the poster event and represented SRL at the career fair – here’s to a wonderful start to the 2015-2016 academic year!
