This week, we focused on learning about convolutional neural networks (CNNs) and examining code examples written in CNN. We reached a point in our code development where we need to design a CNN model for our project. We utilized the resources mentioned in the sources section to gain a comprehensive understanding of CNNs. We covered fundamental topics such as setting up a basic CNN framework and the parameters associated with this framework. Our primary objective for next week is to write a concrete CNN model code and evaluate the accuracy of various CNN models and parameters. Additionally, we delved into comparisons of popular CNN models, including RESNET50, RESNET150, and VGG16.

Learning Objectives

• Understand the fundamental concepts of CNNs
• Explore the process of setting up a basic CNN framework
• Identify the parameters involved in CNN architecture
• Analyze comparisons of popular CNN models

Research Activities

• Conducted a thorough review of online resources and tutorials to gain a solid understanding of CNNs
• Investigated various CNN code examples to familiarize ourselves with practical applications
• Examined comparisons of popular CNN models to evaluate their strengths and weaknesses

Key Takeaways

• CNNs are powerful tools for image recognition and classification tasks
• CNNs employ convolutional layers to extract features from input images
• Pooling layers are utilized in CNNs to reduce dimensionality and enhance computational efficiency
• CNN models are characterized by various parameters, including the number of layers, filter sizes, and activation functions
• Popular CNN models such as resnet50, resnet150, and vgg16 exhibit impressive performance in image classification tasks

Sources

• A Comparison Between the VGG16, VGG19 and ResNet50 Architecture Frameworks for Classication of Normal and CLAHE Processed Medical Images: https://assets.researchsquare.com/files/rs-2863523/v1_covered_2fc7b07e-512b-49a2-8ba7-6e2fed1af2d0.pdf?c=1682958585
• Convolutional Neural Networks (CNNs): https://www.tensorflow.org/tutorials/images/cnn
• A Beginner’s Guide to Convolutional Neural Networks with Implementation in Python: https://www.analyticsvidhya.com/blog/2021/08/a-hands-on-guide-to-build-your-first-convolutional-neural-network-model/
• How to build CNN in TensorFlow: examples, code and notebooks: https://www.tensorflow.org/tutorials/images/cnn
• A Simple CNN Model Beginner Guide !!!!!! | Kaggle: https://www.kaggle.com/code/pavansanagapati/a-simple-cnn-model-beginner-guide
• Convolutional Neural Network Tutorial – Kaggle: https://www.kaggle.com/code/kanncaa1/convolutional-neural-network-cnn-tutorial
• VGG16 vs ResNet vs Inception: A Deep Dive into Image Classification Models: https://towardsdatascience.com/architecture-comparison-of-alexnet-vggnet-resnet-inception-densenet-beb8b116866d

This week we made significant progress in understanding CNNs and their applications. We are in a good position to design and implement a CNN model for our project and evaluate its performance. We will continue our research and development work in the coming weeks.