01-05 December 2025
INCOIS, Hyderabad, India.
| Abstract Submission No. | ABS-01-0261 |
| Title of Abstract | An AI-Framework for Rapid Identification of Coastal and Endangered Species |
| Authors | Bineesh K K, Praveen Rozario J, Shyam KJ*, Harigovind R, Mathews Varghese, Anupama Jims, Felix M Philip |
| Organisation | Chinmaya Vishwa Vidyapeeth |
| Address | Marine Biology Regional Centre, Zoological Survey of India,130, Santhome high road Chennai, Tamil Nadu, India Pincode: 600028 E-mail: kkbineesh@gmail.com |
| Country | India |
| Presentation | Poster |
| Abstract | The accurate identification of coastal and endangered species is fundamental to understanding marine ecosystem health and guiding conservation policy, but traditional survey methods relying on expert knowledge are often a bottleneck for large-scale, rapid biodiversity assessment. To address this, we developed a novel AI-powered framework for real-time species identification, centered on a Convolutional Neural Network (CNN) optimized for deployment on accessible platforms. The model was trained on a comprehensive, annotated dataset featuring a wide range of marine and coastal fauna, and was enhanced with advanced data augmentation and transfer learning techniques to ensure robust performance under variable field conditions, such as underwater imaging and partial visibility. Achieving high accuracy, the framework provides instantaneous classifications, significantly improving the speed and accessibility of species identification compared to manual methods. This research validates the power of deploying advanced AI to bridge the gap between scientific data collection and real-time ecological monitoring, creating a powerful tool for researchers, conservationists, and citizen scientists to support the preservation of biodiversity within the Indian Ocean region and beyond. |
| Are you part of IIOE-2 endorsed project | no |
| Keywords | Species identification, AI-based analysis, coastal biodiversity, marine ecology, Convolutional Neural Networks (CNNs), citizen science, automated detection, conservation technology |
| For Awards | no |