A Literature Review on Real-Time Image Classification for Dragonfly Species Using TensorFlow.js and Biodiversity Monitoring.

EasyChair Preprint 15981

Abstract

In this paper we will explore TensorFlow.js, a JavaScript library for using machine learning in browsers. The main goal is to compare and classify different species of dragonfly based on their visual characteristics by using machine learning models and this paper prepares the ground. The study investigates the effectiveness of TensorFlow.js for analyzing and classifying images in real time and shows the potential of Artificial Intelligence to assist in ecological studies, biodiversity conservation, and entomological research. Based on that classification, images can be stored in a back-end application, by having a store confirmation button in the front-end application. The system should use a camera to capture images, classify them using a convolutional neural network (CNN) model, and store the classified images. The performance of the system will be evaluated based on accuracy, speed, and scalability.

Keyphrases: CNN, Entomological research, TensorFlow.js, biodiversity conservation, machine learning

@booklet{EasyChair:15981,
  author    = {Orion Liçi and Inva Bilo and Karafil Kareçi and Ana Ktona},
  title     = {A Literature Review on Real-Time Image Classification for Dragonfly Species Using TensorFlow.js and Biodiversity Monitoring.},
  howpublished = {EasyChair Preprint 15981},
  year      = {EasyChair, 2025}}