Whale and Dolphin Classification

Aim:

        To classify whale and dolphin species from images using feature extraction with VGG16 and a Support Vector Machine (SVM) model.

Abstract:

       Whale and dolphin species classification is crucial for marine biology, aiding in research and conservation efforts. This project aims to create an image classification system that identifies various whale and dolphin species. Using a parquet file containing images and corresponding labels, the dataset is pre-processed to ensure consistency and quality. The VGG16 deep convolutional neural network is used to extract image features, providing a robust foundation for feature engineering. These extracted features are used to train a Support Vector Machine (SVM) model, known for its effectiveness in classification tasks. The model is then deployed on Streamlit, upload images for classification. This approach combines deep learning with traditional machine learning to create an efficient, user-friendly system that can accurately classify whale and dolphin species. By providing this tool, we aim to support marine biologists, researchers, and conservationists in their efforts to study and protect these important marine mammals.

Introduction:

      Whales and dolphins are among the most captivating marine mammals, known for their intelligence, social behaviour, and unique characteristics. Accurate species classification plays a vital role in marine biology, as it informs research, conservation efforts, and environmental policies. Traditional methods for species classification often require expert knowledge and manual inspection, which can be time-consuming and prone to human error.

     In recent years, artificial intelligence (AI) and machine learning have made significant strides in automating complex tasks. Image classification, in particular, has benefited from deep learning techniques that can recognize intricate patterns in large datasets. The goal of this project is to leverage these advancements to develop a system that can classify whale and dolphin species from images.

      The project uses a combination of deep learning and machine learning techniques to achieve accurate classification. VGG16, a deep convolutional neural network, is used to extract features from the images, providing a comprehensive representation of the visual data. These features are then used to train a Support Vector Machine (SVM) model, known for its robustness and accuracy in classification tasks.

       The deployment of this system on Streamlit allows for a user-friendly interface, enabling users to upload images and receive classification results quickly. This streamlined approach has the potential to significantly reduce the time and effort required for species classification, making it an invaluable tool for marine biologists, researchers, and conservationists.

Existing Method:

      Traditional methods for classifying whale and dolphin species often involve manual inspection and expert knowledge. These methods can be subjective and time-consuming, with varying degrees of accuracy. Some existing automated systems use basic machine learning techniques, but they may lack the accuracy and robustness of deep learning-based approaches.

Proposed Method:

       The proposed method involves a multi-step process to classify whale and dolphin species from images. First, the dataset is collected and pre-processed to ensure high-quality input data. The VGG16 model is used to extract features from the images, capturing complex patterns and details. These features are then used to train a Support Vector Machine (SVM) model, which excels in binary and multi-class classification tasks. The trained model is deployed on Streamlit, providing a simple user interface where users can upload images and get classification results. This combination of deep learning and SVM creates a robust and accurate system for species classification.