Object_Classification_From_Scratch

Create an classification model from scratch on Oxford IIT Data Set

Objective of this notebook

• The purpose of this notebook is to build a image classification on the given data set
• Details of the problem statement , data set , summary of the code/solution , sample output/Prediction from the program and final result of the project are listed in the sections to follow.

Problem Statement

Build a model to classify pet

Data Description:

• Data Set :Oxford IIT Pet Data , a 37 category data set
• Link : https://www.robots.ox.ac.uk/~vgg/data/pets/

Summary of the Solution/Code:

The code aims at building a image classification model

• We begin by building a utility class to to read the input images ,convert them to a structured format for further processing and model building and pickle the created input data files Code @ Read_& Structure_Input.ipynb
• We then do an EDA on the data and check various statistics on the images such as total no of images , whether classes are balanced and display a few images with their bounding boxes for visulaization . Code @ EDA.ipynb
• We will then we will pre-process the input data to make it compatible for model building wherein we first do a Train-Test split and create 2 subsets (70:30 split), convert all images to tensors for both subsets,resize all the images to one pre-defined size & visualize data and see that everything looks good. Code @ Pre-Processing_Data.ipynb
• Finally we build the model, train the model , log the results and run an inference.We build an image classifier.We will try 3 different variants of the model The first variant is we will build a Convolution Neural Network from SCRATCH ,train it & capture results. The second variant is a model which a pre-trained MobileNet base with a basic top layer re-built ,re-trained and log results.The third variant is a model with pre-trained MobileNet base with a slightly more sophisticated top layer re-built, re-trained and log results.At the end of this program , we will have trained model with scores logged for all the different models tried and tested.We compare and contrast score , pick the best model and run an inference. Code @ Train_And_Inference.ipynb

Sample Ouput/Prediction :

Here is a sample Ouput image from the program/model

Result

• 3 Different models were tried to build the classification model
• A model with pre-trained mobile Net base with a basic top layer reconstructed gave the best results with an accuracy of .9936 in 7 epochs
• A model with pre-trained mobile Net base with a more complex top layer reconstructed gave us a result of .8739 in 7 epochs . Further training could possible give us higher numbers because accuracy was steadily increasing with each epoch
• A 6 layer deep CNN model built from scratch gave us a poor result of 0.097 in 7 epochs and was stagnating at that score. This clearly shows us the importance of using CNN architectures created by the pioneers of vision with transfer learning enabled.
• The best model is model no 1( model with pre-trained mobile Net base with a basic top layer reconstructed)

References & Guidance

• PGP GL/UAT study material
• https://colab.research.google.com/github/zaidalyafeai/Notebooks/blob/master/Localizer.ipynb