This month we're hosting Yoav Feldman. Yoav is an experienced R&D in these fields - Computer Vision, Speech Processing ( Speech Recognition & Speaker Recognition ) using Machine Learning approach.
Medium and Large companies and organizations around the globe are using products that were developed by Yoav:
Singapore army, TAP ( Portugal Airline company ), The Netherlands Train Company and more.
Here is a clip of Hand Detection & Tracking that was developed by Yoav : https://www.youtube.com/watch?v=gqiZR7-2ND8
Yoav will present an approach for face detection in live video using a non-patented approach without pre-training the algorithm.
Face detection in live video is a pre procedure to Face recognition or Face tracking, For example, Tracking the face in order to move the mouse cursor.
There is a popular algorithm for face detection that was invented by the two researchers, Viola and Jones, and is a standard in the open source OpenCV and in Matlab.
It has 2 main drawbacks:
- It's patented
- A huge amount of time required to create characteristics for new objects.
In this MeetUp Yoav shall present a solution that will be based on standard Image Processing and Computer Vision Methods.
This solution operates instantly, meaning, it doesn't need any kind of pre machine learning training.
Bonus Lecture - (How) Can software quality be improved?
Yechiel Kimchi is a mathematician (PhD in abstract set theory, Hebrew University,
Jerusalem), a computer scientist (teaching more than fifteen years at the CS faculty
at The Technion, Israel), and a software developer — working more than twenty
years alternating between big high-tech companies and working as a consultant
through his own small firm. Initially in C, and then in C++, he is interested
in object orientation and the ways to develop software that is correct,
maintainable, and efficient at the same time. Among other things, he developed
heuristics for efficiently solving practical NP-hard problems, but he considers
as his greatest achievement the influence he had on the technical education
of several thousand Israeli software engineers.
(How) Can Software Quality be Improved?
Software quality is not about features, it is about stability, bug count,
and alike. It's a bit like the weather: Many talk about it, but very little
seem to be changing. I'll present my view that the problems are not local
(i.e., each company has its own problems/mistakes), but rather are global
(i.e., same kind of errors are very common). Instead of reviewing Mega-LOC
I have reviewed some well known and acclaimed Coding-Standard documents.
My findings are divided to three groups: Missing important coding principles,
containing correct requirements that do not belong in Coding-Standards,
and, surprisingly enough, some wrong principles and misleading observations.
Nevertheless, even after identifying (part of) the problems, major obstacles
stand in the way for quality improvement, and I'll review some of them.