Discriminative histogram taxonomy features for snake species identification

Abstract

Background: Incorrect snake identification from the observable visual traits is a major reason for death resulting from snake bites in tropics. So far no automatic classification method has been proposed to distinguish snakes by deciphering the taxonomy features of snake for the two major species of snakes i.e. Elapidae and Viperidae. We identify 38 different taxonomically relevant features to develop the Snake database from 490 sample images of Naja Naja (Spectacled cobra), 193 sample images of Ophiophagus Hannah (King cobra), 88 images of Bungarus caeruleus (Common krait), 304 sample images of Daboia russelii (Russell's viper), 116 images of Echis carinatus (Saw scaled viper) and 108 images of Hypnale hypnale (Hump Nosed Pit Viper). Results: Snake identification performances with 13 different types of classifiers and 12 attribute elevator demonstrate that 15 out of 38 taxonomically relevant features are enough for snake identification. Interestingly, these features were almost equally distributed from the logical grouping of top, side and body views of snake images, and the features from the bottom view of snakes had the least role in the snake identification. Conclusion: We find that only few of the taxonomically relevant snake features are useful in the process of snake identification. These discriminant features are essential to improve the accuracy of snake identification and classification. The presented study indicate that automated snake identification is useful for practical applications such as in medical diagnosis, conservation studies and surveys by interdisciplinary practitioners with little expertise in snake taxonomy.

A Novel Method for Functional Annotation Prediction Based on Combination of Classification Methods

“…As shown above Figure 1 , IPR has the matrix of many binary features of 2,624 when viewed based on one GO. It is the well-known fact that learning and experimenting by selecting only meaningful features would reduce the time to be taken and have a better result as compared with learning and experimenting the method presented above by these many matrices [ 16 , 21 – 23 ]. When representing the case in which “1” represents that each IPR has protein by positive data and the case of not having it by “0,” the positive negative data is to be counted for each protein.…”

Snake species identification by using natural language processing

“…To our knowledge, the closest work to our research can be found in [10] and [11]. In these works, automatic snake species identification techniques from snake images were proposed by using machine learning algorithms.…”

“…In these works, automatic snake species identification techniques from snake images were proposed by using machine learning algorithms. Amir et al [10] applied texture based approach as features, while James et al [11] used features describing top, side and body views of snake images. In contrast, NLP was utilized in our work to enable snake species recognition through text-based information from a human.…”

Output Effect Evaluation Based on Input Features in Neural Incremental Attribute Learning for Better Classification Performance

“…Many previous studies have shown that preprocessing, like feature selection, feature ordering and feature extraction, usually plays a very important role in the final performance [19][20][21]. Feature ordering is naturally treated as an independent preprocessing stage in IAL [14], because features should be imported into an IAL predictive system one by one.…”