Tech Tutorial: Image segmentation
summary
A hands-on pipeline uses TensorFlow to load images and masks, apply data augmentation, train a convolutional segmentation model and visualize predicted masks overlaid on inputs. The workflow emphasizes a U-Net-like architecture and unit‑level design choices, shows loading ~a few hundred training images, demonstrates mask import/update and validation steps, and packages a runnable demo for visualization (including a lightweight mobile/Android export path) so viewers see the model actually produce pixelwise segmentations.
AI-written summary · about
chapters (6)
- 0:00Introduction and welcome
- 12:42What is image segmentation
- 22:18TensorFlow basics for segmentation
- 30:30Building segmentation models
- 38:19Improving segmentation results
- 42:02Q&A and wrap-up
related talks
01h 29mComputer Vision: An (Un?)Expected Journey, with Keras and Tensorflow
01h 15mTutorial on Image Classification using Scikit-Image, Scikit-learn, and PyTorch
15mTinySeg: Model Optimizing Framework for Image Segmentation on Tiny Embedded Systems
01h 06mDeep Learning with TensorFlow
01h 29mHuman-like Visual Search Application with Small Data
33mWhat and where? - Machine learning for geospatial image analysis
36mDeep Learning - Segmentation, Identification and Classification
51mA Practical Introduction to Convnets for Image Classification