-
Neural Network Trainning
-
DataSet
-
Trainning Tool
-
Image Capture from Dataset
-
Human Machin Interface
-
Identify Botoon
-
That's it
-
Image Captura from camera
-
Swordfish test
-
Salmon test
-
Carp test
-
Catfis
-
Test element that are not in dataset
-
Without Perturbation
-
With more perturbation
-
With large disturbance
-
The Challenge
-
Solution Laser Reference Pattern
-
With small perturbation
-
Camera calibration
-
Result Width and height
-
Testing the app in real time
-
Artificial intelligence
-
It's me
-
Bioinspiration
It is an attempt the imitation of the models and systems of nature for the purpose of solving complex human problems.
Employs Artificial Intelligent (artificial neural networks) , trained with images of a data set , to automatically identify the species of marine fish. With this application tourists, divers and fishermen can become researchers to contribute information that allows to have elements to make management decisions and environmental sustainability.
It also measures the length of the fish in real time with a video camera and a laser pointer.
"...a function that helps non-experts to identify species correctly would be extremely valuable." Fishackathon 2016 Problem Statements (page 3)
Log in or sign up for Devpost to join the conversation.