NAME: Oyster Volume Estimation Data
TYPE: Calibration experiment
SIZE: 30 observations, 5 variables
DESCRIPTIVE ABSTRACT:
The data set contains the results of a calibration experiment designed to estimate volume of
oysters and to compare two computer vision systems (2-D vs. 3-D) for classification of oysters
based on their image size in number of pixels.
SOURCES:
These data were collected in 2001 in a research and development lab at AGRI-TECH Inc., Woodstock,
VA, for calibrating an oyster volume estimation system. Information on how the data was collected
can be found in "Three-dimension Reconstruction for High-speed Volume Measurement," Proceedings of
the International Society for Optical Engineering, 2001. A similar dataset was used in this paper.
VARIABLE DESCRIPTIONS:
Datasets: 30oysters.dat (20oysters.dat is a random sample from 30oyster.dat)
Columns Variables
1- 2 Oyster ID
11-15 Oyster weight (g)
27-31 Oyster volume (cc)
44-50 Oyster size information from the 3-D imaging system (in volume pixels)
54-58 Oyster size information from the 2-D imaging system (in pixels)
Values are aligned and delimited by blanks. There are no missing values.
STORY BEHIND THE DATA:
Shucking and grading of oysters is a very labor intensive and tedious task. An automated oyster
meat grading system integrated with the shucking and packing equipment can totally eliminate human
handling of oyster meat, save cost, and increase the accuracy of oyster grading and shucking. For
a request from a client in oyster packing industry, a group of engineers and scientists was
attempting to design a more efficient 3-D computer vision measurement system for oyster
classification to improve on the existing 2-D system for oyster sorting. Oysters are defined to be
small, medium, or large if their volume falls in (0,10], [10, 13), [13, inf), respectively. Data
on 30 oysters (10 small, 10 medium, and 10 large) were collected in a calibration experiment in
which simple linear regression models were used to predict observed oyster volume, one using
digital image area (2-D) in pixels as a predictor, one using digital image volume (3-D).
PEDAGOGICAL NOTES:
These data can be used to demonstrate simple linear regression with a very clean linear trend,
and to establish the calibration equation for estimating the volume of oyster. Through the
process of volume estimation modeling, statistics can be generated for comparing the
performance of the 2-D and 3-D systems. There are many criteria can be used, such as
prediction errors, coefficient of determination, and the percentage of correct classification.
Other topics include calibration experiments, inverse regression, regression diagnostics, and
regression through the origin.
REFERENCE:
Lee, D., Lane, R., and Chang, G., "Three-dimension Reconstruction for High-speed Volume
Measurement," Proceedings of the International Society for Optical Engineering, Machine
Vision and Three-Dimensional Imaging Systems for Inspection and Metrology, Volume 4189,
p.258-267, 2001.
SUBMITTED BY:
G. Andy Chang
Department of Mathematics and Statistics
Youngstown State University
Youngstown, OH 44555
gchang@ysu.edu
G. Jay Kerns
Department of Mathematics and Statistics
Youngstown State University
Youngstown, OH 44555
gkerns@ysu.edu
D. J. Lee
Department of Electrical and Computer Engineering
Brigham Young University
Provo, UT 84602
djlee@ee.byu.edu
Gary L. Stanek
Department of Mathematics and Statistics
Youngstown State University
Youngstown, OH 44555
stanek@math.ysu.edu