Automated identification and retrieval of moth images with semantically related visual attributes on the wings

How automated image analysis techniques help scientists in species identification and classification?

“…Since SVM is a binary classifier and for classification of multi-class problem it has to use one over all classification for each class, SVM performs better than ANN. Comparing SVM and ANN results with SRV attributes in an automated identification system for moths [25], this classifier outperforms both SVM and ANN classifiers. In [49], classification was based on LBP and the accuracy rate in identification depends on variables such as neighbouring and radius values.…”

“…They used area, perimeter, diameter, compatibility, compactness and solidity as geometrical features, uniformity, median, entropy, variance, inertia, homogeneity and co-occurrence as texture features and Hu1 [69] and Ami1-Ami2 [26] as morphological features. Feng and Bhanu [25] developed a system which adopted semantically related visual (SRV) attributes. They claimed that shape, texture and colour may fail in validity if the images are visually complex and have semantic contents.…”

“…Most of the classification methods are mentioned in [60,89,112], including structural, fuzzy, transform, neural network-based methods and many more. Some automated identification systems [55] employ neural networks or learning algorithms when there are many classes and small Automatic insect classification 10 SVM > 90 [56] Automated identification and retrieval of moth images 50 SRV 85 [25] Automatic identification of species 740 ANN 91-93 [35] Water monitoring -automated and real time identification and classification of algae 23 ANN: SOM 98 [17] Automatic identification of butterfly species 5 ANN 98 [49] Automated system for malaria parasite identification 2 SVM 80 [88] Automatic plant species identification 8 Sparse representation 76-79 [42] Automated identification of copepods 8 ANN 93.13 [55] Automated identification and retrieval of moth images 50 SRV attributes 34-70 [24] Automatic wild animal identification 26…”

Automated Identification of Moth Species

“…Beside mentioned semi-automated systems, there are fully automated identification systems also developed for example Digital Automated Identification System (DAISY) [18], SPecies IDentification Automated (SPIDA) [25,26], Automated Bee Identification System (ABIS) [27], Automate Identification Sarinder Kaur Dhillon, Y.M. Leong, Elham Y.K, Lim L.H.S University Malaya Malaysia Of Bees [19], Automatic Identification Of Whiteflies, Aphids And Thrips [20], Automatic Identification Of Live Moths [21], Automated Insect Identification [22,23], Automated Real-Time Dynamic Identification Of Flying And Resting Butterfly [24]. DAISY was previously tested on various types of organism such as British bumble bees [28], Costa Rican Hawkmoths [8] and British Lepidoptera: Moths [29] while SPIDA was tested mainly on Australian ground spiders [25], ABIS was tested on bees [27].…”