Deep Learning & Computer Vision
Amazon Machine Learning University (MLU)
Amazon Machine Learning University (MLU)
The Machine Learning University (MLU) is an education initiative to help Amazon tech employees gain machine learning (ML) skills. MLU has the goal to significantly accelerate the propagation of ML knowledge beyond our specialized ML scientists. As ML becomes a core part of Amazon’s voice, vision, data, robotics and decision systems, the demand for ML knowledge has outpaced our capacity to hire this highly sought after talent. MLU seeks to bridge this gap organically by having our ML scientists teach our tech employees ML skills.
Instruction team
Instruction team
Mohamed Elgendy
Mohamed Elgendy
Here are a few, quick facts about me:
- Computer vision engineer for the past 5 years at Google, Intel, and Twilio
- Author of Deep Learning for Computer Vision book (publication date December 2018)
- Won the Marketplace-MLU computer vision challenge in June 2018
Gurumurthy Swaminathan
Gurumurthy Swaminathan
Here are a few, quick facts about me:
- I have a PhD in computational neuroscience working on developing human vision models
- I have been working in CV industry for the past 13 years at Honeywell Research and Amazon on various CV problems (e.g., face/people detection, multi-camera tracking, face recognition, image segmentation)
- Currently I work in AWS AI platforms team on Sagemaker CV product offering
Prerequisites
Prerequisites
Please review the following prereqs before the class starts:
- Basic understanding of Python 2 or 3
- Numpy, Pandas, and Matplotlib
- Understand deep learning and CNNs (preferred)
- Set up your dev environment
Curriculum
Curriculum
- Introduction to CV and image classification
- Intro to CV course
- Traditional CV features
- Image classification
- Object detection
- Image segmentation
- Video processing
- 3D processing and GAN
- Hackathon
Week 1
- Intro to Computer Vision
- Computer Vision Pipeline
- Pre-processing
- Selecting areas of interest
- Image segmentation
- Features and Object Recognition
- Convolutional Neural Networks
- Convolutional Layers
- Stride and padding
- Pooling
- Putting it all together using Keras and Tensorflow
- Homework: CIFAR10 image classification