Coral reef fish detection and recognition in underwater videos by supervised machine learning : Comparison between Deep Learning and HOG+SVM methods

Sébastien Villon (1) , Marc Chaumont (2) , Gérard Subsol (2) , Sébastien Villéger (3) , Thomas Claverie (3) , David Mouillot (3)

1 UM - Université de Montpellier
2 ICAR - Image & Interaction
3 UMR MARBEC - MARine Biodiversity Exploitation and Conservation

In this paper, we present two supervised machine learning methods to automatically detect and recognize coral reef fishes in underwater HD videos. The first method relies on a traditional two-step approach: extraction of HOG features and use of a SVM classifier. The second method is based on Deep Learning. We compare the results of the two methods on real data and discuss their strengths and weaknesses.

Domaines: Vision par ordinateur et reconnaissance de formes [cs.CV]; Réseau de neurones [cs.NE]; Traitement du signal et de l'image [eess.SP]; Apprentissage [cs.LG]

Sébastien Villon, Marc Chaumont, Gérard Subsol, Sébastien Villéger, Thomas Claverie, et al.. Coral reef fish detection and recognition in underwater videos by supervised machine learning : Comparison between Deep Learning and HOG+SVM methods. ACIVS: Advanced Concepts for Intelligent Vision Systems, Oct 2016, Lecce, Italy. ⟨hal-01374123⟩