A Multiagent System for Edge Detection and Continuity Perception on Fish Otolith Images
- PDF / 1,867,564 Bytes
- 8 Pages / 600 x 792 pts Page_size
- 34 Downloads / 156 Views
A Multiagent System for Edge Detection and Continuity Perception on Fish Otolith Images Anne Guillaud ´ Ecole Nationale d’Ing´enieurs de Brest, Technopˆole Brest-Iroise, BP 30815, 29608 Brest Cedex, France Email: [email protected]
Herve Troadec Institut Franc¸ais de Recherche pour l’Exploitation des Mers, Center de Brest, Laboratoire de Scl´erochronologie des Animaux Aquatiques, BP 70, 29280 Plouzan´e, France
Abdesslam Benzinou ´ Ecole Nationale d’Ing´enieurs de Brest, Technopˆole Brest-Iroise, BP 30815, 29608 Brest Cedex, France Email: [email protected]
Jean Le Bihan ´ Ecole Nationale d’Ing´enieurs de Brest, Technopˆole Brest-Iroise, BP 30815, 29608 Brest Cedex, France Email: jean.le [email protected]
Vincent Rodin ´ Ecole Nationale d’Ing´enieurs de Brest, Technopˆole Brest-Iroise, BP 30815, 29608 Brest Cedex, France Email: [email protected] Received 5 July 2001 and in revised form 26 January 2002 We present an algorithm for fish otolith growth ring detection using a multiagent system. Up to now, the identification of growth rings, for age estimation, is routinely achieved by human readers, but this task is tedious and depends on the reader subjectivity. One of the major problems encountered during an automatic contour detection is the lack of ring continuity perception. We present an approach to improve this continuity perception based on a 2D reconstruction of rings using a multiagent system. The originality of the approach is to use local edge detection achieved by agents and combine it with continuity perception that active contours allow. Keywords and phrases: edge detection, multiagent system, otolith.
1.
INTRODUCTION
The growth of the otolith is an accretionary process. The otolith structure is made of alternative opaque and translucent concentric rings. The purpose of growth rings identification is to acquire data on age and growth of fish population. Such data are needed in a great number of biological and ecological studies and to improve stock management. Up to now, this analysis has been mainly limited to a ring count. Ring continuity is a major concept on which readers base their ring detection. This paper presents an approach to this continuity perception based on the 2D reconstruction of rings. In 1996 Rodin et al. [1] tried to reconstruct the rings in
polar coordinates using a graph construction, by connecting nodes obtained with a primary segmentation of the image. Most of rings were detected on young individuals, but problems were encountered with older individuals, which have very thin marginal rings. In 1997, Benzinou et al. [2] applied a deformable model (Locally Deformable B-Bubble Model) to otolith images. The model was initialized at the growth center of the otolith and then inflated by computing local forces based on the gray levels and on the global shape of the otolith. Results were encouraging but sometimes drifting of points from the inflated shape could be observed. To tackle this problem, another method inspired by [3] has been proposed in [4]. The external otolith edge is used
Data Loading...
