Course Overview
How does the human visual system derive rich information about the three-dimensional world from the continually changing two-dimensional images that enter the eyes? How can we replicate this capability in computer vision systems? CS332 explores models for analyzing digital images to infer information such as the three-dimensional structure, movement, and color of objects in a scene, and to recognize the objects being viewed. You will see how an interdisciplinary approach to the study of vision, which combines insights from computer science, neuroscience, cognitive science, and psychology, contributes to the design of effective computer vision systems and the understanding of human visual processing. Specific topics include edge detection, stereopsis, motion analysis, color, face recognition, object recognition, and image processing applications in areas such as medicine, security, information retrieval, robotics, assistive technologies, and intelligent vehicles. The course uses vision software written in MATLAB.
Prerequisites: CS112 or CS230 or permission of instructor. No prior
experience with MATLAB is needed.
Distribution Credit: Mathematical Modeling
Units: 1.0
Learning Goals
The aims of this course are to help you develop background and skills that enable you to:
- understand, design, apply, and critically evaluate computational methods for deriving important information about the world from digital images
- draw connections between observations from empirical studies of visual processing in biological systems and the behavior of computational models of vision
- communicate, orally and in writing, technical content related to human and computer vision
- assess the benefits to society, as well as controversial aspects, of applications of computer vision systems in areas such as medicine, security, and assistive technologies
Learning Activities
This course will follow a hybrid flipped/lecture classroom style. In advance of some classes (Tuesdays), there will be required videos, reading, and a short assessment to submit the evening before class. The videos will introduce new concepts and provide details of technical methods and other background information needed for assignment work. There will typically be a total of about 45-60 minutes of video to view each week, divided into shorter segments. For the readings, there is no required textbook for this course — notes, book excerpts, and research articles will be posted online. The assessment that will be posted for some classes will consist of short-answer questions on the new material from videos and readings, submitted via a shared Google doc.
The Tuesday/Wednesday classes will typically include a short recap of material from the videos and reading, time for discussion and questions from the class, and small group activities. Friday classes will meet in the computer classroom, SCI E101, and will usually include work that will be integrated into assignments that will be posted bi-weekly. There will be five homework assignments that combine computer work with written analysis or reflection.
Finally, there will be a final project that involves the independent exploration of a vision topic that is of interest to you. Topics can be explored through current literature that may include journal articles, excerpts from books, or online papers, and can include an implementation or computer simulations with a vision model, or a perceptual experiment. Some possible areas of exploration are listed on the final project web page. During the last three classes of the semester, students will give an 8-10 minute presentation of their final project work. A final paper on this work is due by the end of final exam period, on Thursday, December 22.
There will be no exams in this course.
Grading
All of the above learning activities are required, and count toward your final grade for the course. More specifically, your grade will be based on the following components:
| 15% | Class preparation (videos, readings, assessments) |
| 50% | Assignments (five, roughly bi-weekly) |
| 25% | Final project |
| 10% | Class participation |
Please let me know if you anticipate an absence from class.
Observance of Religious Holidays
Please also let me know if you anticipate an absence, or need other accommodations, for observance of religious holidays. The College provides an online Religious Observance Notification Form that you can use to inform me of this need.
Late Assignment Policy
All assignments are due on the posted due date. Unless otherwise stated, you and your partner can use a 48-hour extension on assignment work, no questions asked. However, if this extra time is not sufficient for you to complete the assignment, you should contact me and discuss a plan for finishing your assignment work. We will work together to make sure that the plan is reasonable and effective so that it supports both your learning and your health. If the solutions to an assignment are distributed before you have turned in your work, you are bound by the Honor Code not to examine them.
The short assessments on the videos and reading that will be posted for Tuesday classes will be due by 8:00 pm on Monday evening before class. This requirement helps to ensure that you are prepared for in-class discussions and work on the new material. If you are not able to complete the assessment by the due date/time, please contact me as soon as possible, before class.
Collaboration Policy
Computer work for assignments will be started during our Friday classes, usually in pairs, and you should complete this computer work with the same partner. The partnerships will change for each assignment. You are welcome to discuss the assignments with students outside your pair, and exchange general ideas about how to solve problems. However, each pair should write their own code and compose their own written summary or analysis of their computer work. For problems requiring only a written analysis or solution, you can discuss high-level strategies for solving the problems with classmates, but each pair should write out the details of the problem solution on their own, and in their own words. It is best if partners can compose this written work collaboratively, but I understand that if you need to work remotely, this may sometimes pose a challenge. It is unacceptable and consititutes a violation of the Honor Code (1) to compose a written problem solution together with students from another group and turn in two copies of the same work, (2) to copy written work from students outside your group, or (3) view solutions from previous terms of CS332.
In keeping with the standards of the scientific community, you must give credit where credit is due. If you make use of an idea that was developed by (or jointly with) others, please reference them appropriately in your work. It is unacceptable for students to work together and not acknowledge each other in their write-ups. If you have questions about this policy, please ask me.
The short assessments on the videos and reading posted for Tuesday classes must be completed individually. They will generally be "open-book," with each assessment providing details on which sources can be consulted.
Course Directory and Student Accounts on the CS File Server
Course materials will be linked from the Schedule & Handouts
page and/or will be available on the CS file server, in the directory /home/cs332. Most
software to be downloaded will be accessible through a zip file linked from the schedule page
and also placed in the /home/cs332/download folder on the server.
Course webpages can be viewed from a web browser at the following URL:
http://cs.wellesley.edu/~cs332/.
Videos can be viewed on YouTube or downloaded from a shared Google folder.
Each student will have an individual account on the CS file server. If you have taken other computer science courses at Wellesley, you probably already have an account. If not, please submit the Server Account Request form to obtain a new account. You can also use this form to change your password.
Course Software
All of the CS332 course software is written in MATLAB. At Wellesley, MATLAB 8 is available on campus computers, including Macs and PCs, and can be installed on your personal Mac or PC for use on or off campus. (The version that you download from the LTS site will be newer than the version installed on campus computers, but our software should be compatible across versions.) For more information about installing MATLAB on your laptop, go to this LTS MATLAB installation page linked from the Wellesley Technology Support webpage and the CS332 documentation page. This platform road map at the MathWorks website provides useful information about the compatibility between MATLAB releases and operating system releases on various platforms. The MATLAB software is key-served, so there are a limited number of copies that can be used at one time, but we do not anticipate a problem with this limit. A student version of MATLAB can be purchased at the Mathworks website for $99.
Specific course software can be downloaded and uploaded to and from personal computers using Cyberduck or other file transfer program such as Fetch (for Macs) or WinSCP (for PCs). As noted above, most software will also be available in zip files linked from the Schedule & Handouts page.
Google Group
A Google Group has been created, CS-332-01-FA21, and students enrolled in this class have been added as members. If you did not receive a confirmation e-mail, please let me know so that I can add you. You can send e-mail to this group at the address CS-332-01-FA21@wellesley.edu. I will post important announcements here, and I encourage you to post questions or comments related to course material or activities, or other topics of interest to students in the class. Feel free to respond to questions posted by other students. Please do not post MATLAB code here — homework discussions should be at a high-level English description.
Disabilities and Accommodations
If you have special needs of any kind, please communicate with me about accommodations that may be helpful to you. If you have a disability or condition, either long-term or temporary, and need reasonable academic adjustments in this course, please contact Disability Services as early as possible in the term so that they can review your request and send a message to me outlining your need for accommodations. If you are unsure but suspect you have an undocumented need for accommodations, you are encouraged to contact Disability Services. They can be reached at disabilityservices@wellesley.edu, at 781-283-2434, by scheduling an appointment online at their website www.wellesley.edu/adr, or by visiting their offices on the 3rd floor of Clapp Library, rooms 316 and 315.
Faculty Responsibilities on Disclosures of Discrimination, Harassment, and Sexual Misconduct
Wellesley College considers diversity essential to educational excellence, and we are committed to being a community in which each member thrives. The College does not allow discrimination or harassment based on race, color, sex, gender identity or expression, sexual orientation, ethnic or national origin or ancestry, physical or mental disability, pregnancy or any other protected status under applicable local, state, or federal law.
Pursuant to Wellesley College policy, all employees, including faculty, are considered responsible employees. This means that any disclosure of discrimination, harassment, or sexual misconduct to a faculty member will need to be shared with the College's Director of Non-Discrimination Initiatives / Title IX and ADA / Section 504 Coordinator. Students who do not wish to have these issues disclosed to the College should speak with confidential resources who are the only offices at the College that do not have this same reporting obligation. Confidential resources include Health Services (781-283-2810, available 24/7), the Stone Center Counseling Services (781-283-2839, available 24/7) and the Office of Religious and Spiritual Life (781-283-2685). You should assume that any person employed on campus outside of these three confidential offices has an obligation to share information with Wellesley College through the Office of Non-Discrimination Initiatives.