Computer Vision: An (Un?)Expected Journey, with Keras and Tensorflow
summary
Hands-on work uses Keras and TensorFlow notebooks (Colab-ready) with concrete pipelines: train classifiers on Fashion-MNIST (simple dense nets hit ~70% then tuned conv nets reach ~92–93%), apply transfer learning with VGG16/ImageNet weights, run TensorFlow Object Detection API to build a pizza detector (per-instance boxes often 90%+), and run Mask R-CNN for pixelwise segmentation (usable masks, occasional failures on crowded scenes). Slides, code, and dataset-prep scripts are linked in the GitHub materials.
AI-written summary · about
chapters (13)
- 0:00Welcome and setup
- 1:55Computer vision techniques
- 3:06Introduction
- 4:10Neural networks history
- 7:03Keras and TensorFlow
- 12:57Simple classification model
- 27:59CNN for image classification
- 33:05CNN architectures with Quo Vadis
- 42:05Transfer learning for classification
- 54:05Object detection with TF API
- 1:02:30Custom object detector training
- 1:18:43Image segmentation with Mask R-CNN
- 1:28:45Summary
related talks
01h 07mComputer Vision with PyTorch
01h 27mHands-On Computer Vision with PyTorch
59mA Friendly Introduction to Deep Learning for Computer Vision
45mTech Tutorial: Image segmentation
01h 01mA practical guide to deep learning
01h 39mGoogle Colaboratory and Python: Deep Learning basics for CV tasks
47mDeep Learning in Computer Vision
33mComputer Vision, Not Just for Breakfast Anymore