RevaWiki

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.


====== Document Indexation and Multimedia Retrieval (USTH mTU3.2) ======

===== General Informations =====
Your Instructor: 
Ass. Prof. Axel Carlier, University of Toulouse
 --- //[[Axel.Carlier@enseeiht.fr|Ass. Prof. Axel CARLIER, University of Toulouse]] //


===== Lecture #1 =====

== Lecture time: ==

Monday Nov 28, 2016, 5:15pm-9:30pm Hanoi Time 

General introduction to the course and the topic.

A few (already-known) administrative points can be found in the subsequent slides.

== Slides ==

----

{{:public:res-ens:interactive-media:l1.1generalintromm2.1-2016.pdf| First part: General Introduction }} 

{{:public:res-ens:interactive-media:exampletoan.pdf| Your first contrib (see also this example of short student Intro)}}

----

{{:public:res-ens:dimr:brainstorming.pdf|}}


----

A few pictures to try existing Content-Based Image Retrieval websites:

{{:public:res-ens:dimr:requete10.jpg?linkonly|}}
{{:public:res-ens:dimr:requete9.jpg?linkonly|}}
{{:public:res-ens:dimr:requete8.jpg?linkonly|}}
{{:public:res-ens:dimr:requete6.jpg?linkonly|}}
{{:public:res-ens:dimr:requete3.jpg?linkonly|}}


----

Slides of the first class:

{{:public:res-ens:dimr:summary_class1.pdf|}}





===== Lecture #2 =====

== Lecture time: ==

Tuesday Nov 29, 2016, 5:15pm-9:30pm Hanoi Time 

Reminder on retrieval: boolean and vector models, evaluation

{{:public:res-ens:dimr:mtu3.2-_2.pdf|}}

At the end of this lecture, you should have started thinking about your own dataset.




===== Lecture #3 =====

== Lecture time: ==

Wednesday Nov 30, 2016, 5:15pm-9:30pm Hanoi Time 

Image retrieval with local features

{{:public:res-ens:dimr:mtu3.2-_3.pdf|}}

At the end of this lecture, you should have started thinking about your own dataset.


===== Lecture #4 =====

== Lecture time: ==

Thursday Dec 01, 2016, 5:15pm-9:30pm Hanoi Time 

Today is all about labs. 

To start the labs, you will need to download the following archive:
[[http://carlier.perso.enseeiht.fr/lire-teaching.zip|Code base for the labs]]

In order to make the code base work you need to:

  * Install gradle
  * Install Java 1.8

Once you have installed these softwares, unpack the archive and run //gradle wrapper//
in the extracted folder.
Open the build.gradle file and change the paths to your test set and one of your query images.
You can test the two sample codes that are provided:

  * ./gradlew runIndexing (on Unix) builds an index of your test set
  * ./gradlew runSearch (on Unix) ranks the test set images according to the similarity to your your query.

Once the code runs, you can start implementing Precision/Recall measures to evaluate the quality of the retrieval. Use your Ground Truth to obtain the Precision/Recall on all query images and compute a global score of Precision and Recall for this algorithm, on all queries.


===== Lecture #5 =====

== Lecture time: ==

Friday Dec 02, 2016, 5:15pm-9:30pm Hanoi Time 

Global descriptors, Bag of Words

{{:public:res-ens:dimr:mtu3.2-_5.pdf|}}

In the mean time, we will continue the labs.