الفهرس 
chapter 1:IntroductionOnly 14 pages are availabe for public view
 |  | from | 97 |  |  |
| | | from | 97 | | |
Abstract
Computer vision is one of the areas that have shown rapid progress. Grand View Research has estimated the global computer vision market to be valued at $11.32 billion in 2020 and is expected to expand in the next few years(Research, Sep, 2021). Therefore, due to the recent advances in computer vision in many industries.
It is important to check where it all began and where it is heading, especially when it comes to choosing your next computer vision project. In this thesis, we’ll cover the foundations and trends in image classification, and image retrieval. And study the evolution of machine learning and deep learning techniques to perform image recognition
In this thesis, we first explore the performance of traditional image features classification method Bag of Visual Words (BoVW), transfer learning methods, an ensemble CNN (Convolutional Neural Network) as feature extractor with dimension reduction technique known as PCA (Principal Component Analysis) classified with SVM (Support Vector Machines), Logistic Regression , and KNN (K-Nearest Neighbor). Then compare these methods with automatic machine learning technique. Experiments have been conducted for evaluating the performance of such methods with various configurations of CNN Among different architectures that we have considered in image classification for small dataset (Bulk Food Grains dataset).
Keywords: Image classification, Image retrieval, Visual feature extraction, Bag of visual words, Deep learning Transfer learning.