$4\mathrm{.}0$ Problem Description

You are working on a computer vision project where you need to develop an AI system to recognize and localize objects in images. You are required to find a computer vision dataset containing images of various objects from different perspectives and backgrounds. Your task is applying Convolutional Neural Networks $($CNN$)$ for image recognition.

$4\mathrm{.}1$ Dataset

This assignment must involve at least ONE dataset. For this assignment, it is recommended to use an existing dataset rather than creating one dataset. It is a good idea to explore publicly available datasets, like those found on UCI or Kaggle. This helps you focus on learning CV techniques without worrying about data collection. Try choosing a less commonly used dataset as this allows for more unique and innovative approaches to the field of CV$.$

$4\mathrm{.}2$ Final Report

Final report

The result of your project is a report that clearly and concisely describes what you did, the results you obtained, and what they mean. Maximum of $20$ pages.

Minimum report requirements

a$)$ Introduction:

$($Total: $5$ marks$)$

$-$ Introduction of the study background $($of the topic chosen$),$ research goal and objectives.

b$)$ Methodology:

$($Total: $\backslash (\backslash$mathbf$\{45\}\backslash )$ marks$)$

Architecture Design $(15$ marks$)$

$-$ Design a CNN architecture for image recognition. Specify the number of convolutional layers, pooling layers, and fully connected layers.

$-$ Justify your design choices based on the nature of the dataset and the complexity of the classification task.

Implementation $(25$ marks$)$

$-$ the designed CNN model using a deep learning framework of your choice $($e$.$g$.,$ TensorFlow, PyTorch$).$

$-$ Include code snippets demonstrating the model architecture, layer configurations, and any necessary preprocessing steps.