computer vision

Computer Vision

Overview

90% of all the data in the world has been created in the last two years. Moreover, 2.5 quintillion bytes of data are created every day.

An inherent challenge lies in how to make sense of all that data. This is where Data Science & A.I. can help out. It can strengthen your business by allowing you to activate your data. With that in mind, Anchormen is organizing value-packed trainings for your employees; we want you to join the future.

The Computer Vision specialization will provide your team with the essentials of Data Science based on real-life cases. Duration of the training is one day and takes place in one of our offices (Amsterdam, Groningen) or at your location.

Artificial Neural Networks

The goal of Artificial Neural Networks is to mimic the human brain. Just like our brain, the neural network learns patterns over time, which are used to compute the best solutions. A neural network uses neurons as components to which data is fed. Neurons work together to solve a problem; for example, recognizing faces or speech. The neural network then alters the pattern in terms of connections between the neurons to change their interaction. This leads to the neural network trying to solve the problem again. Over time, the network learns the best pattern to arrive to the optimum solution. In order to do that, the algorithm breaks down into features. For example, take a look at the picture below:

Herein, different parts of the picture are cut up in smaller pieces. These pieces are used as features for the network to train on and learn from. With enough input, the network will be able to understand the difference between a squirrel and a cat. The approach to recognizing vehicles is similar.

Anchormen has worked on and participated in a couple of interesting cases using Computer Vision. By analyzing pictures and training a neural network, we’ve been able to automate radiology image analysis for classifying tumors and fractures. Also, we’ve analysed which pictures lead to a higher conversion rate when posted on a leisure website. These are examples of some possibilities that are possible with the technology.

Apart from analyzing and predicting, computer vision is also used for creating new images. This has been done as part of “The Next Rembrandt” (https://www.nextrembrandt.com/). Such techniques are part of the creative A.I. field, but leverage the power of Computer Vision as well.

Is this interesting for your organization? The whole package costs a total of €799 per person and is coordinated for groups of 3-10 people.

Contents

The precise content of the training is decided based on a consultation with us. Every training is specially tailored based on the company’s needs and current level of subject matter understanding. Here are some of the topics you can expect to learn about:

  • Background on the basics of computer vision.
  • Feature detection, description and matching methods.
  • Object detection.
  • Object recognition.
  • Category recognition.
  • Practical tips for working on your own computer vision problem.

Prerequisites

Some background in Machine Learning and Python is needed to follow the lectures.

Tools

Python will be the core programming language for assignments, where we rely on the Keras library (backed by either Tensorflow or Theano) to train neural networks. If your graphical card allows it, we use the CUDA library to speed up the computations by running them on the GPU.

Are you interested in this training?

Please, contact our Sales Executive for a preliminary consultation.

Mick van Bakel
+31 6 23 72 96 97
+31 20 773 19 72
m.vanbakel@anchormen.nl