Visual Search using Deep learning (CNN)

Prerequisite :

• You must have python 3.7 or more installed on your system.

• You must have Spyder, jupyter notebook or pycharm installed on your system. Spyder or jupyter notebook come up with anaconda. you just need to launch them after installing anaconda.

• If you work on a colab no need to install python or Any other IDE, you need to just sign in with google colab.

• A Kaggle account is a must to import the data from kaggle.

• Google colab to build this project.

Skilled required:

• Python programming language

• Machine learning concept

• Deep learning concept

What you’ll learn

• How to import data from kaggle on colab

• How to read the image data

• How to build the convolutional neural network (CNN) model

• How to trained and Evaluate the model

• How to compare two image using cosine_similarity

Problem Statement or Description:

Visual search aims at searching for images by visual features to provide users with relevant image lists. In this project build the visual search model using the image dataset which will be able to search similar images and provide to the user.

Key highlights of projects or Essence

• This project is about Visual search using deep learning.

• This project shows you how to use data directly from Kaggle.

• This project shows you how to read the image data and visualize it.

• This project show you how to use image data to train the model

• This project shows you how to compare the similarity of two images using cosine_similarity.

• This project shows you how to display a similar image using this model.

Packages and module used :

• os

• Matplotlib

• Numpy

• Pickle

• cosine_similarity

• Seaborn

• Keras

Recommended projects:

1. Face mask detection

2. Image classification with CIFAR-10 Dataset

3. Dogs breed identification

4. Traffic sign classification

5. Breast cancer detection