%0 Conference Proceedings
%4 sid.inpe.br/sibgrapi/2016/09.02.10.31
%2 sid.inpe.br/sibgrapi/2016/09.02.10.31.39
%T Using Color for Fish Species Classification
%D 2016
%A Freitas, Ueliton,
%A Gonçalves, Wesley Nunes,
%A Matsubara, Edson Takashi,
%A Sabino, Jose,
%A Borth, Marcelo Rafael,
%A Pistori, Hemerson,
%@affiliation Federal Univivesity of Mato Grosso do Sul (campus Campo Grande)
%@affiliation Federal Univivesity of Mato Grosso do Sul (campus Ponta Porã)
%@affiliation Federal Univivesity of Mato Grosso do Sul (campus Campo Grande)
%@affiliation Anhaguera University - Uniderp
%@affiliation Federal Institute of Parana - IFPR
%@affiliation Dom Bosco Catholic University - UCDB
%E Aliaga, Daniel G.,
%E Davis, Larry S.,
%E Farias, Ricardo C.,
%E Fernandes, Leandro A. F.,
%E Gibson, Stuart J.,
%E Giraldi, Gilson A.,
%E Gois, João Paulo,
%E Maciel, Anderson,
%E Menotti, David,
%E Miranda, Paulo A. V.,
%E Musse, Soraia,
%E Namikawa, Laercio,
%E Pamplona, Mauricio,
%E Papa, João Paulo,
%E Santos, Jefersson dos,
%E Schwartz, William Robson,
%E Thomaz, Carlos E.,
%B Conference on Graphics, Patterns and Images, 29 (SIBGRAPI)
%C São José dos Campos, SP, Brazil
%8 4-7 Oct. 2016
%I Sociedade Brasileira de Computação
%J Porto Alegre
%S Proceedings
%K Fish Species Classification, Bag of Visual Words, Color images, Computer Vision.
%X This paper presents an application dedicated to mobile devices whose objective is to classify fish species, using images and concepts of Computer Vision and Artificial Intelligence. The application was developed to Android smartphones with the help of OpenCV Computer Vision library for classification and training phases. The techniques employed in the description of the images are based on Bag of Visual Words applied to color images. They are: HSV and RGB color histograms, Bag of Visual Words, Bag of Features and Colors, Bag of Colors and Bag of Colored Words (BoCW). For the species classification, three types of classifiers was used: Support Vector Machine (SVM), Decision Tree and K-Nearest Neighbors algorithm (KNN). In the experiments several parameters for all the classifiers were tested in order to find the best results for classification. To compare the performance of the feature extraction techniques, as well as the classifiers, the metrics F-Score were used as the main metric and the Area Under the Curve (AUC) as an auxiliary metric. The technique with best result was BoC using the SVM classifier.
%@language en
%3 Paper.pdf