Learn how to make your drawings come alive… Lecture 4: SKETCH RECOGNITION Analysis, implementation, and comparison of sketch recognition algorithms, including feature-based, vision-based, geometry-based, and timing-based recognition algorithms; examination of methods to combine results from various algorithms to improve recognition using AI techniques, such as graphical models. Learn how to make your drawings come alive…
Class Overview Long Paper discussion (lead by ?) Brief Lecture on Long Discussion of previous homework and implementation issues MARQS: Presentation and Discussion Homework
Long Paper Discussion: (lead by ?) Thoughts?
Class Overview Long Paper discussion (lead by ?) Brief Lecture on Long Discussion of previous homework and implementation issues MARQS: Presentation and Discussion Homework
Class Overview Long Paper discussion (lead by ?) Brief Lecture on Long Discussion of previous homework and implementation issues MARQS: Presentation and Discussion Homework
Homework Implementation Discussion Compare recognition results. What did people get? Implementation difficulties?
Performing Recognition on these Features What do you think about the features? What recognition problems did you have? Can you think of better features? (this is your homework)
Class Overview Long Paper discussion (lead by ?) Brief Lecture on Long Discussion of previous homework and implementation issues MARQS: Presentation and Discussion Homework
Class Overview Long Paper discussion (lead by ?) Brief Lecture on Long Discussion of previous homework and implementation issues MARQS: Presentation and Discussion Homework
Homework Read Paulson paper. Invent and better recognition features Write a paper on the results.
Paper Criteria CONTENT: FORMAT Clarity Ideas Abstract Previous work Implementation/Methodology Results Discussion Future Work Conclusion Bibliography
Paper Criteria: Content Clarity There should be no spelling mistakes There should be no grammar mistakes Writing should be clear and cogent Writing should be as brief as possible, i.e., not rambling Uses images when necessary to clarify ideas Ideas Ideas are original Ideas are interesting
Paper Criteria: Format/Abstract A one paragraph summary of your paper One sentence motivation One sentence what you did One sentence results Do not make this a story People should be able to *only* read your abstract and know exactly what you did
Paper Criteria: Format/Introduction Introduction/Motivation Introduce the area Describe the problem you are trying to solve Why is this problem important Provide any background information necessary to understand the problem Any intelligent person should be able to understand, and be motivated by, your problem
Paper Criteria: Format/Previous Work List at least 2-3 of the most related work in the field (at a minimum you should have Rubine/2 Long papers/Paulson) Describe how your work differs from theirs (i.e. why their work does not solve the problem you are trying to solve)
Paper Criteria:Format / Implementation-Methodology What did you do? How did you do it? How can other people reproduce what you did?
Paper Criteria: Format/Results What was the outcome of you work? What statistical test did you use to determine these results (i.e., a t-test / recognition accuracy) Graphs and tables Make sure there is some way to measure what you have done
Paper Criteria: Format/Discussion Analysis of the results What worked What didn’t work Why do you think things worked Why do you think things didn’t work
Paper Criteria: Format/Future Work If you had more time to work on this, what would you do next give the results of your paper? Why? What future ideas naturally extend from this work?
Paper Criteria: Format/Conclusion Summarize what you told them What were the key findings Similar to abstract, but you can assume people have read the paper What did you want people to get out of the paper? What should they walk away remembering?
Paper Criteria: Format/Bibliography References are properly cited Refer to papers read for proper format.