Metadata augmented deep neural networks for wild animal classification

TL;DR

This paper presents a novel wildlife classification method that combines image data with metadata, improving accuracy and robustness in challenging conditions, and demonstrating high performance even with metadata alone.

Contribution

It introduces a metadata-augmented deep neural network for wild animal classification, enhancing accuracy and reducing dependence on image quality.

Findings

Accuracy increased from 98.4% to 98.9% with metadata integration

Metadata-only classification achieved high accuracy

Method outperforms existing image-only approaches

Abstract

Camera trap imagery has become an invaluable asset in contemporary wildlife surveillance, enabling researchers to observe and investigate the behaviors of wild animals. While existing methods rely solely on image data for classification, this may not suffice in cases of suboptimal animal angles, lighting, or image quality. This study introduces a novel approach that enhances wild animal classification by combining specific metadata (temperature, location, time, etc) with image data. Using a dataset focused on the Norwegian climate, our models show an accuracy increase from 98.4% to 98.9% compared to existing methods. Notably, our approach also achieves high accuracy with metadata-only classification, highlighting its potential to reduce reliance on image quality. This work paves the way for integrated systems that advance wildlife classification technology.