Automated annotation of corals in natural scene images using multiple texture representations

Abstract

Current coral reef health monitoring programs rely on biodiversity data obtained through the acquisition and annotation of underwater photographs. Manual annotation of these photographs is a necessary step, but has become problematic due to the high volume of images and the high cost of human resources. While automated and reliable multi-spectral annotation methods exist, coral reef images are often limited to visible light, which makes automation difficult. Much of the previous work has focused on popular texture recognition methods, but the results remain unsatisfactory when compared to human performance for the same task. In this work, we present an improved automatic method for coral image annotation that yields consistent accuracy improvements over existing methods. Our method builds on previous work by combining multiple feature representations. We demonstrate that the aggregation of multiple methods outperforms any single method. Furthermore, our proposed system requires virtually no parameter tuning, and supports rejection for improved results. Firstly, the complex texture diversity of corals is handled by combining multiple feature representations: local binary patterns, hue and opponent angle histograms, textons, and deep convolutional activation feature. Secondly, these multiple representations are aggregated using a score-level fusion of multiple support vector machines. Thirdly, rejection can optionally be applied to enhance classification results, and allows efficient semi-supervised image annotation in collaboration with human experts.

A diver-operated hyperspectral imaging and topographic surveying system for automated mapping of benthic habitats

“…Our analyses show that hyperspectral imagery is conducive to automated mapping of the reef benthos with a high degree of taxonomic detail, while requiring very little manual annotation. The annotation effort can be further reduced by adopting techniques from recent efforts in computer vision research 23 .…”

“…Such analysis typically entails sub-sampling the high-resolution images and visual identification by experts to estimate benthic coverage 18 , 19 . Recent developments have improved algorithmic automation and structuring of the annotation effort, through projects such as CATAMI and CoralNet, to alleviate the backlog of field survey analysis 20 – 23 . Other efforts have explored the use of spectral filtering to estimate benthic coverage 24 , 25 .…”

Computer Vision and Deep Learning for Fish Classification in Underwater Habitats: A Survey

“…Since it is difficult to obtain a large labelled dataset, various techniques have been proposed to address this challenge. Some of the techniques applied to the fish and marine habitat monitoring domains include transfer learning [89], data augmentation [30,48], using hybrid features [90][91][92], weakly supervised learning [93], and active learning [94].…”

Coral Reef Image Classifications with Hybrid Methods

“…Jean-Nicola Blanche [14] presented VGG Net for classification problems. Neural network and nearest neighbour classifiers are used.YANGuoqiang [15] implemented Support Vector Machine (SVM) to monitor the healths of coral reefs.M.Bennamoun [16] presented Resfeats for image classification and object detection.The deep layers of CNN is used to give best performance in classification.AniBrownMary [17] proposed Improved Local Derivative Pattern (ILDP) for classification.…”