Computer Vision Project Directory

This directory is a collection of computer vision projects developed by me, Carlos Eduardo Fontaneli, during my studies. The goal is to showcase my skills and knowledge in image processing, shape recognition, and other computer vision techniques. Each project in this repository demonstrates a different aspect of computer vision and its applications.

Projects

1. Recognizing Shapes in Images with OpenCV

Course Project: This project is part of the "Recognizing Shapes in Images with OpenCV" course I completed. It focuses on applying computer vision techniques to recognize shapes in digital images using OpenCV.

• Skills Demonstrated: Image processing, edge detection, Hough transforms, shape recognition.
• Key Tasks:
  • Loading and displaying digital images.
  • Computing image gradients to highlight edges.
  • Detecting edges using the Canny edge detector.
  • Recognizing lines and circles in images using Hough transforms.

2. Image Classification with TensorFlow

Guided Project: This project is part of the "TensorFlow for Convolutional Neural Networks" series. It aims to reinforce skills and build more projects with TensorFlow, focusing on the practical application of convolutional neural networks (CNNs) for image classification.

• Skills Demonstrated: Convolutional neural networks, data preprocessing, image classification, TensorFlow.

• Key Tasks:

  • Preprocessing training and validation image datasets with data augmentation.
  • Building and training a CNN model for binary image classification (dogs vs. cats).
  • Evaluating the model's performance on the validation set using metrics like accuracy, precision, recall, and F1-score.
  • Visualizing predictions and probabilities for real-world images.

  Aqui está uma seção adicional para o seu README.md sobre o projeto de augmentação de imagens:

3. Image Augmentation Techniques with TensorFlow

Personal Project: This project explores various image augmentation techniques using TensorFlow's ImageDataGenerator class. The goal is to enhance the diversity of the training dataset and improve the generalization of machine learning models.

• Skills Demonstrated: Image augmentation, TensorFlow, data preprocessing.
• Key Tasks:
  • Implementing different augmentation operations such as rotation, width and height shifts, shear, zoom, flips, rescaling, and preprocessing for specific neural network architectures.
  • Applying these augmentations to a training dataset to create a more robust and varied set of training examples.
  • Training a convolutional neural network (CNN) model on the augmented dataset and evaluating its performance.
  • Analyzing the impact of each augmentation technique on the model's ability to generalize and perform on unseen data.