Level: Intermediate
John Salmon, Assistant Professor, Brigham Young University
JMP provides some interesting image processing features that are not often used or showcased. Although images can be stored and presented in classic JMP tables, additional properties and scripting can enhance the capabilities in JMP to process and analyze images. This paper explores some of the image presenting and processing capabilities on a data set of over 4000 images of food. These images of food represent a time-stamped record of everything the author ate during a full calendar year. Leveraging crowdsourcing provided through Amazon's Mechanical Turk website, these images were shared with hundreds of individuals who estimated the caloric content of each image and imported into JMP. The analysis and visualization of this caloric data with corresponding images are compared and validated against the estimates of a trained nutritionist. Additional examples of image processing through JMP scripting is discussed and presented.
