|
The Caffe2 Framework for Mobile and Embedded Deep Learning
Fei Sun, software engineer at Facebook, introduces Caffe2, a new open-source machine learning framework, in this presentation. Sun also explains how Facebook is using Caffe2 to enable computer vision in mobile and embedded devices.
Methods for Understanding How Deep Neural Networks Work
In this presentation, Dr. Wojciech Samek of the Fraunhofer Heinrich Hertz Institute provides an overview of deep neural networks and their capabilities and limitations for visual perception tasks. Among other topics, Dr. Samek covers various approaches for peering inside the network model "black box" and determining how it works, as well as the degree of reality behind widespread popular claims that artificial intelligence systems are or will soon be "unbeatable".
|
|
Using Vision to Collect Rich Data in the Moment of Truth for Retail Analytics and Market Research
Dr. Anja Dieckmann of GfK Verein and Markus Iwanczok of GfK SE explore the value of using vision-based market research, going beyond the capabilities of traditional questionnaires and retail panels, in this presentation. Dieckmann and Iwanczok explain how to infer emotions from facial expressions, as well as how to use in-store tracking to log and boost shopper conversions into sales.
The Reverse Factory: Embedded Vision in High-volume Laboratory Applications
In this presentation, Dr. Patrick Courtney, MBA, of tec-connection and the Standards in Laboratory Automation (SiLA) Consortium, explains how vision-based systems can effectively find use in various high-volume laboratory data collection and analysis scenarios. Courtney describes the underlying motivation and market opportunity, covers a detailed case study example, and provides insights into future imaging trends.
|
