[Pycon] [new paper] "Sandeep Saurabh" - Image Classification using Deep Learning (Capsule Networks)

[info a pycon.it]

Title: Image Classification using Deep Learning (Capsule Networks)
Duration: 45 (includes Q&A)
Q&A Session: 0
Language: en
Type: Talk

Abstract: Capsule Networks is a paradigm shifting technique in deep learning with amazing Computer Vision applications. The talk will explain the idea behind CapsNet and it’s working in detail with example from the research paper I am working on.

I am going to present my talk on the topic of -Image Classification using Deep Learning (Capsule Networks). The talk will explain the comprehensive idea and mathematics behind CapsNet and it’s possible applications with focus on the research paper I am currently working on. I have structured my talk as follow :

0-10 mins - 
1. Introduction to the problem statement and it’s importance in the Computer Vision world 2. Explaining CNN ,why it’s used in Computer Vision and the problems it has solved

10-30 mins - 
1. Analysing CNN further and explaining the disadvantages and shortcomings of CNN
2. Introducing the latest research in the field -Capsule Networks and how it can corrects CNN
3. Explaining mathematics and Computer Graphics aspect of CapsNet
4. Explaining the neural network architecture consisting of CapsNet

30-45 mins -
1. Showing the working and results of Image Classification with CapsNet as per my research paper
2. Comparing the results with CNN based results and explain the advantages

45-50 mins -
1. Further presenting the possible amazing applications possible with CapsNet

50 - 60 mins - Interactive Q&A session with audience

Tags: [u'ComputerVision', u'computer-graphics', u'Artificial Intelligence']