Image Recognition with Keras and TensorFlow
summary
A practical walkthrough of building a convolutional neural network with Keras and TensorFlow to solve an image-recognition task: detecting X-ray orientation (frontal vs. lateral). Covers the full pipeline in a Jupyter notebook, from image preprocessing through training and evaluation, with application context in medical imaging.
AI-written summary · about
chapters (22)
- 0:00Introduction
- 1:54Why Keras and TensorFlow
- 2:38How CNNs work
- 4:28CNN data input
- 6:14Building CNN architecture
- 8:27Compiling and training
- 11:18Model evaluation
- 12:41Limitations of custom CNN
- 13:44Transfer Learning with Latam Gelatti
- 16:15How Transfer Learning works
- 18:22Application in cardiomegaly
- 20:11Data augmentation with ImageDataGenerator
- 21:58Setting up Transfer Learning
- 23:20Weight initialization
- 25:06Model coupling
- 26:11Loss function and metrics
- 27:14Adaptive learning rate
- 28:39Training and checkpoints
- 30:06Performance evaluation
- 31:36Critical attention points
- 34:53Problem definition
- 35:13Q&A
related talks
01h 29mComputer Vision: An (Un?)Expected Journey, with Keras and Tensorflow
51mA Practical Introduction to Convnets for Image Classification
53mNowoczesne uczenie sieci neurownowych w TensorFlow
45mTech Tutorial: Image segmentation
46mImage Recognition with Deep Learning From prototype to production
29mComputer Vision Dog Breed Classification with Convolutional Neural Networks, TensorFlow and Kubeflow
01h 06mDeep Learning with TensorFlow
57mTensorflow, deep learning and modern convolutional neural nets, without a PhD