Scribd
Upload
33 views5 pages

Multiclass Fish Image Classification

The project focuses on multiclass fish image classification using deep learning techniques, specifically training CNNs and leveraging transfer learning with pre-trained models. It includes data preprocessing, model evaluation, and the deployment of a Streamlit application for real-time predictions. Deliverables consist of trained models, an interactive web app, Python scripts, a comparison report, and a well-documented GitHub repository.

Uploaded by

purvaja2606
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
33 views5 pages

Multiclass Fish Image Classification

The project focuses on multiclass fish image classification using deep learning techniques, specifically training CNNs and leveraging transfer learning with pre-trained models. It includes data preprocessing, model evaluation, and the deployment of a Streamlit application for real-time predictions. Deliverables consist of trained models, an interactive web app, Python scripts, a comparison report, and a well-documented GitHub repository.

Uploaded by

purvaja2606
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

Project Title

Multiclass Fish Image Classification

Skills take away From


This Project Deep Learning, Python, TensorFlow/Keras,
Streamlit, Data Preprocessing, Transfer Learning,
Model Evaluation, Visualization, and Model
Deployment.

Domain
Image Classification

Problem Statement:

This project focuses on classifying fish images into multiple categories using deep learning
models. The task involves training a CNN from scratch and leveraging transfer learning with
pre-trained models to enhance performance. The project also includes saving models for later
use and deploying a Streamlit application to predict fish categories from user-uploaded
images.

Business Use Cases:

1.​ Enhanced Accuracy: Determine the best model architecture for fish image
classification.
2.​ Deployment Ready: Create a user-friendly web application for real-time predictions.
3.​ Model Comparison: Evaluate and compare metrics across models to select the most
suitable approach for the task.
Approach:

●​ Data Preprocessing and Augmentation


○​ Rescale images to [0, 1] range.
○​ Apply data augmentation techniques like rotation, zoom, and flipping to
enhance model robustness.
●​ Model Training
○​ Train a CNN model from scratch.
○​ Experiment with five pre-trained models (e.g., VGG16, ResNet50, MobileNet,
InceptionV3, EfficientNetB0).
○​ Fine-tune the pre-trained models on the fish dataset.
○​ Save the trained model (max accuracy model ) in .h5 or .pkl format for future
use.
●​ Model Evaluation
○​ Compare metrics such as accuracy, precision, recall, F1-score, and confusion
matrix across all models.
○​ Visualize training history (accuracy and loss) for each model.
●​ Deployment
○​ Build a Streamlit application to:
■​ Allow users to upload fish images.
■​ Predict and display the fish category.
■​ Provide model confidence scores.
●​ Documentation and Deliverables
○​ Provide comprehensive documentation of the approach, code, and evaluation.
○​ Create a GitHub repository with a detailed README.

Dataset

●​ The dataset consists of images of fish, categorized into folders by species. The dataset
is loaded using TensorFlow's ImageDataGenerator for efficient processing.
●​ Dataset:Data as Zip file

Project Deliverables
1.​ Trained Models: CNN and pre-trained models saved in .h5 or .pkl format.
2.​ Streamlit Application: Interactive web app for real-time predictions.
3.​ Python Scripts: For training, evaluation, and deployment.
4.​ Comparison Report: Metrics and insights from all models.
5.​ GitHub Repository: Well-documented codebase.

Project Guidelines:

●​ Follow coding standards: Consistent naming conventions, modular code.


●​ Data validation: Ensure all data is accurate and complete.

Streamlit Doc https://docs.streamlit.io/library/api-reference

Streamlit recording (Tamil) Special Session for STREAMLIT Tamil

Project Live Evaluation


Project Live Evaluation

How to Use GitHub.pptx


GitHub Reference

Orientation
Project Orientation (English)
Project Evaluation metrics:
●​ Maintainable: It can be maintained, even as your codebase grows.
●​ Portable: It works the same in every environment (operating system)
●​ You have to maintain your code on GitHub.(Mandatory)
●​ You have to keep your GitHub repo public so that anyone can check your
code.(Mandatory)
●​ Proper readme file you have to maintain for any project
development(Mandatory)
●​ You should include basic workflow and execution of the entire project in the
readme file on GitHub
●​ Follow the coding standards: https://www.python.org/dev/peps/pep-0008/
●​ You need to Create a Demo video of your working model and post in
LinkedIn(Mandatory)

PROJECT DOUBT CLARIFICATION SESSION ( PROJECT AND CLASS DOUBTS)

About Session: The Project Doubt Clarification Session is a helpful resource for resolving questions
and concerns about projects and class topics. It provides support in understanding project requirements,
addressing code issues, and clarifying class concepts. The session aims to enhance comprehension and
provide guidance to overcome challenges effectively.
Note: Book the slot at least before 12:00 Pm on the same day

Timing: Monday to Saturday (4:00PM to 5:00PM)

Booking link :https://forms.gle/XC553oSbMJ2Gcfug9

LIVE EVALUATION SESSION (CAPSTONE AND FINAL PROJECT)

About Session: The Live Evaluation Session for Capstone and Final Projects allows participants to
showcase their projects and receive real-time feedback for improvement. It assesses project quality and
provides an opportunity for discussion and evaluation.
Note: This form will Open on Saturday and Sunday Only on Every Week

Timing: Monday-Saturday (5:30PM to 7:00PM)


Booking link : https://forms.gle/1m2Gsro41fLtZurRA

You might also like