https://github.com/lmicovic/neural-network-image-recognition
Application that recognize clothes items from images.
https://github.com/lmicovic/neural-network-image-recognition
fashion-mnist keras machine-learning mathplotlib mnist-classification neural-network numpy opencv pandas python3 tensorflow
Last synced: 24 days ago
JSON representation
Application that recognize clothes items from images.
- Host: GitHub
- URL: https://github.com/lmicovic/neural-network-image-recognition
- Owner: lmicovic
- Created: 2024-12-02T17:48:57.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-12-02T18:57:14.000Z (over 1 year ago)
- Last Synced: 2025-05-22T11:35:50.583Z (11 months ago)
- Topics: fashion-mnist, keras, machine-learning, mathplotlib, mnist-classification, neural-network, numpy, opencv, pandas, python3, tensorflow
- Language: Jupyter Notebook
- Homepage:
- Size: 8.33 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Image Data Analysis
## Project Description
This project focuses on the [FashionMNIST](https://www.kaggle.com/datasets/zalando-research/fashionmnist) dataset, which serves as a more exciting and challenging variant of the well-known [MNIST](https://en.wikipedia.org/wiki/MNIST_database) dataset of handwritten digits.
The dataset consists of grayscale images with dimensions of 28x28 pixels, containing 60,000 training images and 10,000 test images.
### Project Objectives
1. **Neural Network Construction**
The first part of the project requires building a neural network using the [Keras library](https://keras.io/), achieving an accuracy of at least 85% on the [FashionMNIST](https://www.kaggle.com/datasets/zalando-research/fashionmnist) test set. The neural network will be trained on a sample of 10,000 training images for efficiency.
2. **Image Processing and Object Recognition**
The second part involves additional image processing using the [OpenCV library](https://opencv.org/) to recognize multiple clothing items in a single image. Several test cases have been prepared for this purpose, consisting of 512x512 images on a white background (with noise) featuring approximately 10 clothing items. The items may be randomly scaled but are always correctly oriented.
The [neural network](https://en.wikipedia.org/wiki/Neural_network_%28machine_learning%29) identifies all clothing items in each test case, drawing bounding boxes around them in blue. Additionally, the program displays the class of each clothing item in red.
## Requirements
- [Python 3.x](https://www.python.org/downloads/)
- [Keras](https://keras.io/)
- [TensorFlow](https://www.tensorflow.org/)
- [OpenCV](https://opencv.org/)
- [NumPy](https://numpy.org/)
- [Matplotlib](https://matplotlib.org/) (optional, for visualizations)
## Installation
1. Clone this repository:
```bash
git clone https://github.com/lmicovic/neural-network-image-recognition.git
```
2. Navigate to the project directory:
```bash
cd project
```
3. Install the required packages:
```bash
pip install -r requirements.txt
```
4. Verify that dependencies are installed:
```bash
pip list
```