Datasets
Standard Dataset
Nile Tilapia (Oreochromis Niloticus ) images on non controlled illumination conditions
- Citation Author(s):
-
Fernando JoaquínRamírez Coronel
Tecnológico Nacional de México/I.T. de Hermosillo, División de Estudios de Posgrado e Investigación, Hermosillo, Sonora, MexicoEdgardEsquer MirandaUniversidad Estatal de Sonora, Hermosillo, Sonora, MexicoOscar MarioRodríguez ElíasTecnológico Nacional de México/I.T. de Hermosillo, División de Estudios de Posgrado e Investigación, Hermosillo, Sonora, Mexico - Submitted by:
- Fernando Ramirez
- Last updated:
- Tue, 05/16/2023 - 12:20
- DOI:
- 10.21227/kq7f-4h27
- Data Format:
- License:
139 Views- Categories:
- Keywords:
Abstract
Dataset consisting of 17 images of Nile Tilapia bred in a circular pond (diameter=10m, deep = 1.5m ) located at the tronconal in Hermosillo, Sonora, México and an aquarium.
From the total of fish, 17 were selected. Each fish was measured with a ruler; the longitude was measured from the beginning of tail to tip and the height was measured from the beginning of the dorsal fin to the bottom, and its weight was measured with a standard balance.
The zip contains a directory (folder) with the images of 17 Tilapia fish and a .CSV file with the data of the 17 fish.
The data encompasses:
- Tilapia number: the number of fish that corresponds with the number of image in de folder.
- Length (cm): the measured height from tail to tip.
- Height (cm): the height was measured from the beginning of the dorsal fin to the bottom.
- Weight (gm): weight measured with a standard balance.
Machine learning algorithms can be applied to the data in the .CSV as linear regression methods and power models to estimate biomass using a combination of paramaters.
Computer vision methodologies can be applied to the images to estimate parameters based on the data in the .CSV file, which can be used as labels for the images.
A common model to test the data is the linear regression between Length and Weight (L-W relationship).
The results obtained by the regression or machine learning methods solely applied to the data of the .CSV files can be compared against the computer vision methodologies applied to the images and .CSV data.
