Image-based fish recognition

We are studying image-based fish identification. Most of traditional approaches used a fish image which was easy to extract a fish region with a white background or uniform background for automatic processing. This research adapted an approach to give several points by manual operation by the user. The proposed approach is able to accept the fish image in the complicated background taken on the rocky place. Furthermore, to investigate the efficient features for fish recognition, we defined various features, such as, shape features, local features, and six kinds of texture features. We collected 129 species under various photography conditions, and the proposed method was carried out to it. As the results, it was confirmed that a combination features with geometric features and BoVW models obtained the highest recognition accuracy.

[1]  Bill Triggs,et al.  Histograms of oriented gradients for human detection , 2005, 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).

[2]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[3]  Robert M. Haralick,et al.  Textural Features for Image Classification , 1973, IEEE Trans. Syst. Man Cybern..

[4]  Robert B. Fisher,et al.  Underwater Live Fish Recognition Using a Balance-Guaranteed Optimized Tree , 2012, ACCV.

[5]  Frank Storbeck,et al.  Fish species recognition using computer vision and a neural network , 2001 .

[6]  Matti Pietikäinen,et al.  Performance evaluation of texture measures with classification based on Kullback discrimination of distributions , 1994, Proceedings of 12th International Conference on Pattern Recognition.

[7]  Mary M. Galloway,et al.  Texture analysis using gray level run lengths , 1974 .