Learning from Design Heritage
Research Incubation Workshop on Data-driven Methods
* The first class: 11am, Monday, February 3 (Room 8-119)
* Undergrad students should sign up 4.550 and grad students should do 4.570.
* Limited enrollment: 8-12 students
* Spring 2025 class does not include any field trip component
* This class is a sequel following the same subject number in spring 2013-24, and may be taken repeatedly with permission of the instructor.
* Homepage: https://cat2.mit.edu/4.570
Spring 2025 - Class Information
This class investigates recent information and visualization technologies that helps studying "design heritage", spatial designs that surround our lives. Design heritage broadly includes architecture, city and landscape; the built, demolished, and planned; and culturally important as well as the banal ones. We will look at various data-driven methods relevant to learn them, such as image/video feature detection, machine learning, physiological sensors, natural language processing, photogrammetric capture, augmented/virtual reality, and gamification. By examining how to collect data, how to process the raw data into forms useful for evaluation, and how to interpret and apply the findings, the students build a foundation for research projects bettering our understanding of the design heritage around us.
Each week during the first half of the class, the class will invite a guest speaker, conduct a short hands-on exercise on a selected data processing tool, and read relevant literature from previous research projects in design heritage. The second half is run in a workshop format with desk critiques, where students are expected to design and develop a small research project individually or in a group. There is no requirement for computational skills for this class, although familiarity with some scripting language is an advantage.
Instructors:
Takehiko Nagakura (takehiko@mit.edu), Daniel Tsai and weekly guest speakers
TA: Xiaoyun Zhang (xiaoyunz@mit.edu)
Regular Class Hours: Monday 11:00am - 2:00pm, room 8-119
Lab (subject to change): Tuesday, 7:00pm - 8:30pm, room 5-216
* Lab hours are used for demos during the first half of the semester, and for project development by student teams during the second half.
* The required final presentation of this class takes place during the last class just before the MIT exam week in May 2025.
Printable course summary: pdf
Course Schedule (subject to change): pdf
Student Work Examples
- Student final projects from 2024 class (Research prototypes)
- Selected student final projects from 2023 class (Research prototypes)
- Student final projects from 2022 class (Research prototypes)
- digital heritage projects and activities from the previous fieldworks (Representation prototypes)
Assignment links
* The content below are for 2024 class. The updated 2025 version will be made available once the class starts.* Refresh your browser page to get the updated materials as the course resource is updated frequently.
- Exercise 1: Image feature analysis
- Exercise 2:
Simulated experience - collecting data and measuring human response
Eye tracking tools: GazeRecorder web app GazeRecorder desktop app Realeye Platform (new!) - Exercise 3 (as part of Ex2): Photogrammetric modeling and panoramic video in VR
Polycam/Recap Photo/Metashape 360 video cameras - Exercise 4: Video semantics analysis
Colab template 1 - YOLO v11/12 (Detection, Segmentation, Pose, Classification)
Colab template 2 - Moments in Time (Activity and event detection)
Colab template 3 - YOLO world (zero-shot object detection for videos) - Exercise 5: Image-to-image generation + training
pix2pix Colab template (by Xiaoyun) pix2pix Colab template (original)
Meta OpenArt AI Meta OpenArt AI Manual (wiki) - Exercise 6:
Natural Language Processing for Design Studies
addendum: more architectural text lab kit - Final Project Proposal
- Final Presentation and Submission
Reference
- Relevant projects from recent MIT theses and conferene papers
- MET warehouse (no AR)
- MET warehouse (with AR)
| A field-based design research project on Machu Picchu, an example using various technology methods |
