Advanced Topics in Mobile Computer Vision (EECS 598-07) - Fall 2010

Instructor: Prof. Silvio Savarese

Webpage: http://www.eecs.umich.edu/~silvio/

Classroom:3433 EECS
Time: T Th 9:00am-10:30am
Office hours: by appointment

Grader: Fisher Yu (yufu at umich dot edu)

 

 

Course schedule :: Annoucements & Resources :: Android Development :: Projects

Course Description
The course surveys recent developments in computer vision, graphics and image processing (e.g, object recognition and categorization, human activity recognition, scene understanding and modeling, virtual augmention, feature extraction, image enhancement) for mobile applications. As part of the main class project, students will design and implement state-of-the-art techniques on a mobile platform (Google Android) which will be available for each student team.

Motivation
The ability to automatically extract visual concepts (objects, scenes or events) from images or videos is a key ingredient in applications where the search queries for information retrieval are images or videos rather than text or sounds. Moreover, once objects are identified, extracting geometrical attributes (objects pose and 3D shape) is crucial if the image must be augmented with additional virtual elements so as to enhance the viewer experience. For instance, in applications such as Google Goggles, users can point their cell phone to a landmark scene (e.g., the Statue of Liberty), and hope to retrieve useful information (e.g., historical background, business hours, etc...).  In order to do so, algorithms must be designed for enabling either locally (on the device) or externally (on the cloud) the recognition of the dominant object – i.e., the Statue of Liberty - or the semantic content of the scene. Furthermore, as the user observes the scene through the phone, the device may superimpose virtual objects such as an older historical view of the statue. At that end, algorithms must be designed so as to estimate statue’s pose, its rough 3D shape, the scene layout and enable photorealistic superimposition.

Requirements:

  • Present 1-2 set of papers
  • Read papers and participate at class discussion during paper presentations
  • Course project: replicate existing methods or implement new research ideas on the mobile platform.

Grading policy:

  • Class participation & discussion: 15%
  • Paper presentation (quality, clarity, depth, etc.):  30%
  • Course project (quality of the project presentation, work, writing, etc): 55%
  • Project late policy: 25% if one day late; 50% if two days late; zero credit if more than two days

Prerequisites:
Knowledge of linear algebra, probability as well as concepts introduced in 442, or 556 (or equivalent) are necessary for understanding the material covered in this class. Java (or C) programming experience is highly desirable.

Sponsors:
Google kindly donated the Android Platforms which will be used in this course.


***

PAPER PRESENTATIONS:

- Each lecture has a theme. Two students will be presenting a set of papers covering that theme. While preparing your presentation, you must try to integrate the papers as much as possible into a coherent presentation. The presentation will need to summarize the papers and include:
- Introduction: introduce the problem , expain why it is important to solve it; indicate the method that is proposed to solve it.
  - Review of previous work; this is an important session; make sure that an appropriate background is given. Don't hesistate to review previous/preliminary concepts that are critical for the undestanding of the presented work. If a good background is given, it is easier to explain the details of the method and technical solution later on.
  - Why the presented method is better than previous work; and/or explain the key contributions of this work;
  - Technical part: Summary of the technical solution, followed by the details of the technical solution;
  - Experiments: present here experimental results with plots, graphs, images and visualizations.
  - Conclusions: what's the take home message?
Length: The presentation needs to be about 70 minutes long (35 minutes each);Make sure the presentation doesn't go over 70 minutes, so to have time to go through the 5-minute discussion panel.
Evaluation :
- the presentation will be evaluated based on the clarity of the presentation, quality of the slides, how well you get your message across, and how well you handle the questions at the end.
.

Class partecipation:
- Be prepared to ask questions during and after the lecture. At the end of each lecture, two discussion leaders are randomly selected: the discussion leader will ask questions to the presenters and lead a 5-10-minute discussion panel; the quality of the questions & discussion panel will be used for evaluating class participation.

***

PROJECT:

Final Project Report Format
- Max 10 pages;
- Title and authors
- Abstract: short summary of the project with main results
- 6 sections:
  - Sec 1. Introduction: introduce the problem you want to solve, expain why it is important to solve it; and indicate the method you used to solve it. add a concept figure showing the overall idea behind the method you are presenting.
  - sec 2.1. Review of previous work (i.e. previous methods that have explored a similar problem)
  - sec 2.2. Say why your method is better than previous work; and/or summarize the key main contributions of your work;
  - sec 3.1: Technical part: Summary of the technical solution
  - sec 3.2: Technical part: Details of the technical solution; you may want to decompose this section into several subsections; add figures to help your explanation.
  - Sec 4: Experiments: present here experimental results of the method you have implemented with plots, graphs, images and visualizations.
  - Sec 5: Conclusions: what's the take home message?
  - Sec 6: References
- Final format: pdf, please!
You can look at one of my recent publications (such as this or this) as an example.
Evaluation :
- Your project report will be evaluated based on the quality of the writing, the clarity of your technical explanation and, overall, how well you get your message across. If you follow the structure above, you'll have good chances to do a good job. :)


Project Source Code :
There is no need to attach a print out of the source codes to the manuscript. Final source codes of your working program need to be collected into a unique (zipped) file; this file is due on the project submission deadline date and it is supposed to be sent to the grader as indicated by email.

Project Presentation in Class
- The presentation must be 10-15 minutes long; if you have a 2-person team, then each of you will present for 5-7 minutes. There will be 2 minutes for questions after the presentation. If your presentation lasts more than 10-15 minutes, it will be stopped. So please make sure the presentation doesn't go over 10-15 minutes.
Presentation format:
- The idea is to turn your project report (as discussed above) into slides; thus, your presentation will need to include slides covering:
  - Introduction: introduce the problem you want to solve, expain why it is important to solve it; and indicate the method you used to solve it.
  - Review of previous work (i.e. previous methods that have explored a similar problem)
  - Why your method is better than previous work; and/or explain the key main contributions of your work;
  - Technical part: Summary of the technical solution, followed by the details of the technical solution;
  - Experiments: present here experimental results of the method you have implemented with plots, graphs, images and visualizations.
  - Conclusions: what's the take home message?

Evaluation :
- your team will be evaluated based on the clarity of the presentation, quality of the slides, how well you get your message across, and how well you handle the questions at the end. Note that the presentation can still contain ongoing (non final) results; final results may be included in the final report.

Project progress report format
- max 4 pages;
- 3 sections:
  - title and authors
  - sec 1. intro: problem you want to solve and why
  - sec 2. technical part: how do you propose to solve it?
  - sec 3. milestones achieved so far
  - sec 4. remaining milestones (dates and sub-goals)
  - references
- final format: pdf, please!

Project proposal format
- A table where you specify:
-- project title
-- brief description
-- key technical tools
-- Application/impact

Spotlight project presentation
- details will be given in class

Hello world application demo
- details will be give in class