| |
Computational Visual Cognition Laboratory |
Aude Oliva, Ph.D. |
|
Prof. Aude Oliva, Prof. Ki Goosens, & Dr. Sonal Jhaveri Lectures: Friday, Sept. 14, 10:30 - 12:00 pm & Friday, Sept. 21, 11:00 - 12 pm. Location: Building 46-5056 Web page of Graduate Fellowships Web page of Postdoctoral Fellowships Description:This seminar is designed to provide a brief overview of scientific grant writing, and to facilitate application for predoctoral fellowships. We will meet together as group to review some of the fellowships that are most commonly applied for during the first two years of graduate study in Cognitive Science, Computation and Neuroscience, and to discuss specific techniques for scientific writing. These meetings will occur on Friday, September 14 (10:30-12:00pm) and Friday, September 21 (11:00-noon) in room 46-5056. You will be encouraged to search for and identify predoctoral fellowships for which you are eligible, and apply for at least one fellowship this fall. We will be available to meet with you one-on-one during the course of the semester to provide feedback on drafts of all portions of the application (scientific proposal, CV, other essays); you will also solicit feedback from your current lab advisor. For proposals related to Cognitive Science, Computation and Cognitive Neuroscience, contact Aude Oliva (oliva@mit.edu), room 46-4065. For proposals related to Neurosciences/Life Sciences, contact Ki Ann Goosens (kgoosens@mit.edu), room 46-2171B. For questions and advise about grant writing, contact Sonal Jhaveri (sonal@mit.edu), room 46-6023a. |
Prof. Aude Oliva Mondays 5-7pm Department of Brain and Cognitive Sciences, Room 46-4054 Synopsis:We will discuss the state of the art articles about visual scene understanding, in visual cognition, computational vision and cognitive neuroscience. We will review the experimental paradigms, findings and theories with which we can evaluate the capabilities and ?limits of human visual understanding, and we will assess how knowledge about human perception could be used to guide machine vision systems. Topics will include scene gist recognition, space perception, perceptual grouping in complex images, texture perception, navigation, object recognition in clutter, the role of attention, eye movements, and task constraints in understanding the visual world, etc. The seminar will be organized around presentations based on reviews of a few articles per class. |
Instructor: Prof. Aude Oliva (oliva@mit.edu) Monday - Wednesday 10:30 - 12:00 Building 46 - 1015. Description:Teaches principles of experimental methods in human visual perception and cognition, including experimental design, programming (matlab) and statistical analysis. Combines lectures and hands-on experimental exercises; requires three independent experimental projects. To foster improved writing and presentation skills in conducting and criticizing research in visual cognition, students are required to provide reports and give oral presentations of team experiments. Required Textbooks:
|
Instructor: Prof. Aude Oliva (oliva@mit.edu) Mondays 4:30 -7:30p Class Schedule New Graduate Course, Department of Brain and Cognitive Sciences, Room 46-4054 Synopsis:One remarkable aspect of visual recognition is that humans are able to recognize the meaning (or gist) of complex visual scenes effortlessly, with in 1/20 of a second. This rapid understanding of phenomenon can be experienced while looking at rapid sequences in television advertisements and quick cuts in modern movie trailers. How is this remarkable feat accomplished? Taking a cognitive science perspective, the course will detail the experimental paradigms, findings and theories with which we can evaluate the capabilities and limits of human visual understanding, and will assess how knowledge about human cognition could be used to guide machine vision systems. Topics will include scene gist recognition, space perception, perceptual grouping in complex images, texture perception, navigation, memory of events, object recognition in clutter, the role of attention, eye movements, actions and task constraints in understanding the visual world, etc. The course will include:
Requirements:Knowledge of matlab.Note:Participants will work on matlab exercices by team of two. A laptop (PC or Mac) per team will be needed for the class. The course is part of the educational activity of a National science Foundation CAREER award on scene understanding (IIS 0546262), awarded to Prof. Aude Oliva. |
Instructor: Prof. Aude Oliva Monday - Wednesday 10:30 -12:00 pm Building 46 -1015 Description:Teaches principles of experimental methods in human perception and cognition, including experimental design, programming (matlab) and statistical analysis. Combines lectures and hands-on experimental exercises; requires three independent experimental projects. To foster improved writing and presentation skills in conducting and criticizing research in cognitive science, students are required to provide reports and give oral presentations of team experiments. Required Textbooks:
Additional Materials: |
Friday, February 17th- Speakers from a variety of disciplines (neurophysiology, cognitive neuroscience, visual cognition, computational neuroscience and computer vision) will address cutting edge research in visual image and real world scene recognition, detection of objects in natural scenes, as well as the role of attention in scene recognition and visual art. The goal is to encourage exchanges between researchers of all fields of brain sciences in the burgeoning field of scene understanding. Organizers:Aude Oliva, Thomas Serre, Antonio Torralba |
Jan 18-21, 24-26, 12:00 - 1:30pm E25-111 Summary of the class:An introduction to state-of-the-art research on real-world scene comprehension, including the newest developments in human visual cognition, cognitive neuroscience and computational vision. Real-world scenes are 3-dimensional complex visual structures composed of a variety of objects, textures, colors, materials and spatial layouts. Despite this complexity, the human brain understands novel scenes quickly and effortlessly. How is this remarkable feat accomplished? How can we find an object in a cluttered scene? Which strategies does the brain implement to quickly recognize an environment and navigate intelligently? How can we design machine vision systems that are able to recognize complex scenes as humans do? |
Prof. Edward Adelson & Prof. Aude Oliva Wednesdays 4:15 - 5:45 pm Location: NE20-445 This year we will study, in depth, two or three papers that describe models for the analysis or synthesis of texture and other statistical aspects of images. The idea is to take apart the algorithms to understand the issues in vision that the authors confronted and the decisions they made. We will modify the algorithms to understand how they can be pushed in other directions. One example will be Portilla and Simoncelli (Int. Jnl. Comp. Vis., 2000), for which Matlab code is available. Students should have a basic knowledge of Matlab, signal processing, and human and machine vision. |
 |  |  |
| Massachusetts Institute of Technology, 77 Massachusetts Avenue, Building 46-4078, Cambridge, MA 02139 Phone: 617 258 9675 | cvcl.mit@gmail.com |