Image Classification and Captioning Using Deep Neural Networks

Overview

Image captioning is both an exciting and challenging task as it combines two major fields in artificial intelligence: computer vision and natural language processing.

Thus an image captioning system must comprehend both image semantics and natural language. Colloquially, image captioning serves to describe objects in images.

In this case study, we shall use a convolutional neural network (CNN) encoder–recurrent neural network (RNN) decoder method. In this framework, a CNN is used to generate a fixed‐length vector representation, and an RNN to generate a visual description using this vector  representation.

As a component of image captioning, image classification is a classic problem in computer vision, in its own right. A typical image classification task consists in, given an image, identifying it as being a member of one of several fixed classes.

It enters into numerous applications, including autonomous driving, through the critical primitive, namely, fast image classification. Moreover, in modern social media and photo sharing and/or storage applications such as Google Photos make abundant use of image classification techniques to personalise and enhance user experience.

Programme details

9.45/10.00am: Registration

10.00/10.30am: Datasets and pre-processing

11.15/11.45am: Coffee/tea

11.45/12.15am: Optimisation algorithms and regularisation

1.00/1.30pm: Lunch

2.00/2.30pm: Image classification models

3.15/3.45pm: Tea/coffee

3.45/4.15pm: Image captioning models

5.00/5.30pm: Course disperses

Fees

Description Costs
Tuition Fee (includes Tea/Coffee) £100.00
Baguette £5.50

Funding

If you are in receipt of a UK state benefit you may be eligible for a reduction of 50% of tuition fees.

If you do not qualify for the concessionary fee but are experiencing financial hardship, you may still be eligible for financial assistance.

Concessionary fees for short courses

Tutor

Dr Mcebisi Ntleki

Tutor

Dr Ntleki obtained his DPhil in Oxford on a topic in computational biolinguistics. He has taught courses in mathematics, statistics and statistical programming using medium size and large data sets arising in medical research.

Accommodation

Accommodation is not included in the price, but if you wish to stay with us the night before the course, then plesae contact our Residential Centre.

Accommodation in Rewley House - all bedrooms are modern, comfortably furnished and each room has tea and coffee making facilities, Freeview television, and Free WiFi and private bath or shower rooms.  Please contact our Residential Centre on +44 (0) 1865 270362 or email res-ctr@conted.ox.ac.uk for details of availability and discounted prices.