• Subjects 29 to 32 uploaded.


  • The Matlab tool has been updated, both accelerometer and light data from the eWatch are now shown and can be adjusted using the offset using the Control GUI (Download).


  • The Matlab tool has been updated, motion capture can now be displayed without a timestamp file which will be created automatically (mocapTime.txt) although the motion capture data will not be synchronized (Download).

  • Motion capture files from subject 13 have been converted to AMC and ASF files to allow the visualization in the matlab application (View).

  • Some changes in the "how to load data.pdf" (View).
  • Home

    The CMU Multi-Modal Activity Database (CMU-MMAC) database contains multimodal measures of the human activity of subjects performing the tasks involved in cooking and food preparation. The CMU-MMAC database was collected in Carnegie Mellon's Motion Capture Lab. A kitchen was built and to date twenty-five subjects have been recorded cooking five different recipes: brownies, pizza, sandwich, salad, and scrambled eggs.

    The following modalities were recorded:

      • Video:
        (1) Three high spatial resolution (1024 x 768) color video cameras at low temporal resolution (30 Hertz).
        (2) Two low spatial resolution (640 x 480) color video cameras at high temporal resolution (60 Hertz).
        (3) One wearable high spatial resolution (800 x 600 / 1024 x 768) camera at low temporal resolution (30 Hertz).
      • Audio:
        (1) Five balanced microphones.
      • Motion Capture:
        (1) A Vicon motion capture system with 12 infrared MX-40 cameras. Each camera records images at 4 megapixel resolution at 120 Hertz.
      • Internal Measurement Units (IMUs):
        (1) Wired IMUs (3DMGX).
        (2) Bluetooth IMUs (6DOF).
      • Wearable devices:
        (1) BodyMedia.
        (2) eWatch.

      There are 3 datasets:

      • Main Dataset: 43 subjects cooking 5 recipes. View
      • Anomalous Dataset: Three subjects cooking five recipes. Some anomalous situations occur (fire and smoke, falling dishes, distractions...). View
      • Pilot Dataset: First recording system. View

    Technical Report

    Guide to the Carnegie Mellon University Multimodal Activity (CMU-MMAC) Database

    F. De la Torre, J. Hodgins, J. Montano, S. Valcarcel, R. Forcada and J. Macey.

    Tech. report CMU-RI-TR-08-22, Robotics Institute, Carnegie Mellon University, July, 2009.


    This video shows several of the recorded signals while a subject is cooking a salad:

    You can download the visual application used to show all the signals and also the video in AVI format here.

    This video video has been made with the information provided by the Motion Capture, the subject is cooking scramble eggs:

    Download AVI