Underwater fish detection with weak multi-domain supervision

Conference Publication ResearchOnline@JCU
Konovalov, Dmitry A.;Saleh, Alzayat;Bradley, Michael;Sankupellay, Mangalam;Marini, Simone;Sheaves, Marcus
Abstract

Given a sufficiently large training dataset, it is relatively easy to train a modern convolution neural network (CNN) as a required image classifier. However, for the task of fish classification and/or fish detection, if a CNN was trained to detect or classify particular fish species in particular background habitats, the same CNN exhibits much lower accuracy when applied to new/unseen fish species and/or fish habitats. Therefore, in practice, the CNN needs to be continuously fine-tuned to improve its classification accuracy to handle new project-specific fish species or habitats. In this work we present a labelling- efficient method of training a CNN-based fish-detector (the Xception CNN was used as the base) on relatively small numbers (4,000) of project-domain underwater fish/no-fish images from 20 different habitats. Additionally, 17,000 of known negative (that is, missing fish) general-domain (VOC2012) above-water images were used. Two publicly available fish-domain datasets supplied additional 27,000 of above-water and underwater positive/fish im- ages. By using this multi-domain collection of images, the trained Xception-based binary (fish/not-fish) classifier achieved 0.17% false-positives and 0.61% false-negatives on the project’s 20,000 negative and 16,000 positive holdout test images, respectively. The area under the ROC curve (AUC) was 99.94%.

Journal

N/A

Publication Name

2019 International Joint Conference on Neural Networks (IJCNN)

Volume

N/A

ISBN/ISSN

2161-4407

Edition

N/A

Issue

N/A

Pages Count

8

Location

Budapest, Hungary

Publisher

Institute of Electrical and Electronics Engineers

Publisher Url

N/A

Publisher Location

Budapest, Hungary

Publish Date

N/A

Url

N/A

Date

N/A

EISSN

N/A

DOI

10.1109/IJCNN.2019.8851907