Practical Uses of Scala Tensorflow for Image Analysis
summary
Concrete demonstration of using Scala TensorFlow to load TensorFlow protobuf model graphs (~150 MB), feed pixel data as tensors and read outputs (e.g., 19 predictions per sample). Shows notebooks, packaging the code as a JVM service with a REST API, and an image/video pipeline that applies face and license‑plate anonymization via blurring and a simple tracking algorithm; pre‑trained models from public repositories are used and notebooks run locally (live demo failed but code works).
AI-written summary · about
chapters (11)
- 0:00Introduction
- 1:25Talk objectives
- 3:30Machine learning basics
- 15:42TensorFlow overview
- 17:46Jupyter demo
- 22:09Linear model training
- 33:13Deep learning images
- 39:29Object detection service
- 44:23Image anonymisation
- 47:32Summary
- 50:22Q&A
related talks
39mDive into Deep Learning with Scala
01h 56mA Quick Primer on TensorFrames: Apache Spark and TensorFlow Together
22mVue + TensorFlow.js: Building Your First ML Powered App
33mDeploying and Scaling Spark ML and Tensorflow AI Models
38mReal-world Machine Learning with TensorFlow and Cloud AI
37mSimplifying Training Deep & Serving Learning Models...using Tensorflow
20mTensorFlow.js Bringing Machine Learning to the Web and Beyond
42mDeploying ML at Scale with Kubernetes and TFX