Tuesday, December 27, 2005

My PhD resources page

To organize the resources I use in my PhD research, I created a web page to list relevant references. The first version of this page, xml-based, is now available here. I hope it will help me in the future.

Monday, December 26, 2005

New MSc webpage

I created a page for my MSc work at the CG&M website. Currently, it is a preliminary version, submitted for appreciation to my advisor, but the revised final version could be used to define the template for MsC's web pages. This page is available here. Any comments or suggestions will be appreciated.

Thursday, December 22, 2005

First phase of DecoSketch II user testing

Today, I've helped our students at the first phase of user testing within their "DecoSketch II" senior year project.

A testing session.

Monday, December 19, 2005

Released blog for my weekly reports

Finally I found a solution for the weekly reports publishing problem. This solution was a compromise between sending these reports by e-mail to my advisor and publishing them in this blog. I created a new blog just for my weekly reports (in Portuguese).

Sunday, December 18, 2005

Last weekend for Christmas shopping

... for lots of people. Not for me! I'd rather prefer to paddle in the Tagus river. A few photos of this trip are published in a fotos.sapo album.

Kayak trip on Tagus river, this weekend.

Wednesday, December 14, 2005

Correct abstract for WSCG 2006

We have corrected the abstract of our paper accepted at the 14th International Conferences in Central Europe on Computer Graphics, Visualization and Computer Vision (WSCG 2006). After all authors agree with its content, it can be submitted to the conference web site as a final version.

Tuesday, December 13, 2005

Paper acepted at WSCG 2006

Our paper titled "A Calligraphic Interface for Managing Agents" was accepted to be presented as as short communication at the 14th International Conferences in Central Europe on Computer Graphics, Visualization and Computer Vision (WSCG 2006). Now, we need to correct the abstract, according to reviewer's comments, until December 16th. Then, we need to produce a camera-ready version of the paper by January 3rd.

Monday, December 12, 2005

Out of Office

For several reasons, last week I have been out of office. Thus, no real work were done in this period of time.

Friday, December 02, 2005

Converting editION to VC++7.1

After a couple of months on the hold I restarted my work in the editION prototype. However, during this period of time I had updated the C++ compiler from VC++6.0 to VC++7.1. Thus, I needed to make a few changes in the prototype. The more mind-boggling task was related with socket communication. I could not understand, but exactly the same code that before works well, now doesn't. It only started to work well when I forced the system to wait for connection, blocking the GUI. Before it could go on normally and if after a given timeout, the connection could not be established, a error message was displayed.

Short Survey on Tablet PC's

We need to buy a new Tablet PC for our research group. Therefore, I made a short survey in leading brands: Toshiba, Acer and HP. In my humble opinion I think the best offer is the Acer Travelmate C200, but the final decision is not mine...

PhD 2005 yearly report submitted

My PhD 2005 yearly report was submitted to the department server. Next monday I will deliver a printed version of it, signed by my advisor, in the department.

First draft of IR project proposal

The first draft of the project proposal for the Information Retrieval course is ready. Now it will be reviewed in order to produce a more elaborate version of it.

Wednesday, November 30, 2005

Protocol for Task Analysis in DecoSketch

In a near future my PhD work will be used in the DecoSketch project. More precisely, a few months from now a simple 3D shape retrieval mechanism should be integrated in our Computer Aided Decoration prototype. Thus, I am now fully engaged in this project. Today I write the first draft of the protocol for the task analysis. This document is now to be completed in cooperation with the senior year students of this project.

Tuesday, November 29, 2005

Reviewing a paper for WSCG 2006

I've reviewed a paper for the 14th International Conferences in Central Europe on Computer Graphics, Visualization and Computer Vision (WSCG 2006). This paper focused on a hierarchical topological structure for a discrete modeling tool and presents a multi-level approach which provides an effective way to propagate changes in one level of representation to another.

Monday, November 28, 2005

A large video wall

Last Friday we visited the Brisa Operational Control Centre (CCO). Brisa owns most of the Portuguese highways. In their control centre they manage the entire infrastructure. To help them in such task they installed there a large video wall. This device was the main reason for our visit.

View of Brisa's CCO

Friday, November 25, 2005

Restarting work on editION

After a couple of months in stand by, we are going to focus our attentions again on the evolutION platform and its calligraphic interface, editION. This decision were formalized in a meeting between the team members.

PhD 2005 yearly report 2nd version

The second version of my PhD 2005 yearly report is finished [PDF 67KB]. The major changes were made on the work plan and the abstract. Now, if my advisor agrees, I can submit it to the department.

Thursday, November 24, 2005


Regarding my last week post on personal bibliographic/research software, Daniel Gonçalves reminded me of a senior year project he advised two years ago. The BiblioPad were developed to help researches organize bibliographic references. This application extracts information from papers, such as title, authors and references. With this data it establishes relationships among papers, organizing automatically the bibliography.

However, this is not what I am looking for. I want a tool that allows me not only to organize but also classify the bibliographic references I use during my research work. This classification should not only focus on the content of the paper but also other information, such as relevance, source, author credibility, affiliated institutions, related projects, etc...

So, I am still looking for such a tool.

Shape-Based Retrieval and Analysis of 3D Models

Thomas Funkhouser et al published last June a paper, in Communications of the ACM, that introduces the problems in 3D shape analysis and provide a road map for possible solution methods. This paper could be considered extremely relevant for my PhD work. Thus, I made a twenty minutes presentation [ZIP 774KB] about the paper and present it in the Information Retrieval course I am attending.

Monday, November 21, 2005

Following the balls

To understand the problems MultiModal Painter (MMP) senior students are facing I decided to apply some of my computer vision skills trying to solve one of them. To that end I implemented a small program that follows a set of balls in a pre-recorded video. Using MATLAB 6.5, it took me only a few lines of code to achieve it. The resulting video is available for download (ZIP [148KB]).

Snapshot of detection video.

Tuesday, November 15, 2005

Personal Bibliographic/Research Software

In order to maintain a database of my bibliographic citations and research notes I'm searching for a bibliographic/research Software. I have found a some interesting freeware applications, such as Scholar's Aid 4 Lite or BiblioExpress. Does anyone have any suggestions? Would you tell me your personal experience with this or other applications? I’ll be grateful.

PhD work plan v 0.3

After a first iteration with my advisors, I reviewed the work plan. In this new version I expect to deliver the thesis proposal sooner, leaving more time for research, development and testing.

Revised Gantt chart of PhD work plan.
(Click on chart for larger version)

Monday, November 14, 2005

PhD 2005 yearly report

In IST, PhD students must deliver by mid-December an yearly report on the progress of their work. I've just finished the first version [PDF 88KB] of my report. Now I am waiting for feedback from my advisor in order to produce a deliverable version of this document.

Thursday, November 10, 2005

PhD work plan v 0.2

In the beginning of my PhD I devised a preliminary work plan. However, this draft was quite inaccurate and I decided to simply discard it and create a new one from the scratch.

Thus, I have finished the initial version of my PhD work plan, depicted below. Next I’m going to present and discuss it with my advisor and co-advisor in order to produce a plan as accurate as possible at this stage.

Click for larger version
Preliminary Gantt chart of PhD work plan.
(Click on chart for larger version)

Wednesday, November 09, 2005

Co-advisory of PhotoFinder senior project

Since the PhotoFinder senior project is closely related to the work I developed during my master thesis, I'm co-advisoring this senior project. Today I attended a project meeting.

Note: Despite the similar name this project is not directly related with the University of Maryland PhotoFinder.

Tuesday, November 08, 2005

CALIdemo V1.0

The first version of the demonstration tool for CALI is finished. This tool allows the user to check the features extracted by CALI recognition library, as described in our paper "Generic Shape Classification for Retrieval" presented at GREC 2005.

An windows version of this tool is available for download (ZIP [361KB]) as a compressed stand-alone executable file. Installation or configuration is not required, just unzip and run.

Screen-shot of CALIdemo V1.0
Screen-shot of CALIdemo V1.0.

Monday, October 31, 2005

Revised GREC 2005 paper finished

I have just finished the revised version of our paper on "Generic Shape Classification for Retrieval" (original version). This paper will be re-submitted for review for the post-GREC 2005 proceedings, to be published in a Springer LNCS book.

My contribution to Senior Projects report template

In order to help students from IMMI group, we are creating a report template that will not only provide uniformization to senior projects reports but also give students the main topics they should focus on their reports.

This morning I reviewed the latest version of the template proposal and add some contributions to it.

Friday, October 28, 2005

Demonstration Tool for CALI

At IMMI group we use a calligraphic recognition library (CALI), developed a few years ago by Professor Manuel João da Fonseca. Several researchers and students are currently using this library in several applications successfully, some of them with little knowledge of the technique behind the recognizer.

In order to illustrate how CALI works, I am developing a simple demonstrator that lists in detail the features extracted by CALI for each scribble. A preliminary windows version of this application is now available for download (ZIP [379KB]). This is a compressed stand-alone executable file. Installation or configuration is not required, just unzip and run.

Screen-shot of CALIdemo preliminary version.
Screen-shot of CALIdemo preliminary version.

Updating Senior Projects Home Page

Our senior students must create a web page for their senior project, "Trabalho Final de Curso" (TFC) in Portuguese. Moreover, these pages should respect a given template and must be stored in a single place. This is my responsability. Based on the Computer Graphics and Multimedia (CGM) web pages I developed a template for TFC's pages and I am now gathering and publishing these.

Some of these pages are already accessible in a temporary address, while the CGM server aren't up and running.

Fourier Descriptors

After spending some time studying Zernike Moments for shape classification I concluded that to fully understand this technique I must also study the Fourier Descriptors (FD) method. However, it is not as simple as it seems, maybe because I don't remember Fourier Transforms.

I start my research at the CVOnline entry on FD. Does anyone have any other suggestions?

Thursday, October 27, 2005

Backing Up MSDN CD's

Today I need to return the MSDN CD's  with  Visual Studio .NET and libraries to IST computer and network services (CIIST). Thus, I decided to make a backup copy in order to have some indepence from the CIIST when the MSDN disks are necessary to install or update something in my computer, which occurs frequently.

I get some empty CD's and started the copy proccess. However, despite my attempts to write at the maximum recording speed (48×), the writing took always about twenty minutes by CD. When I verified the writing speed I noticed that it were 4×.

During the duplication of the last CD, I started labeling the new discs. Then I discovered why the copy was taking so much time. Instead of CD-R discs, I am using CD-RW discs "compatible 1×, 2× & 4×". Damn!!! How can such detail get unnoticed!?

Conclusion: a task that should take a few minutes, consumed more than one hour plus the time I spent trying to "fix" the problem. At least I made some bibliographic research while backing up the CD's.

Tuesday, October 25, 2005

Zernike Moments...

Today I was supposed to write one or two paragraphs about the use of Zernike moments in shape classification and retrieval, preferably comparing this method with CALI. However, at the end of the day I came out with nothing but a colossal headache. I believe that understanding Zernike moments takes more than one afternoon.

Requesting sponsorship to attend WSCG 2006

Two weeks ago we submitted a paper to the 14th International Conferences in Central Europe on Computer Graphics, Visualization and Computer Vision (WSCG 2006).

Although I do not known yet if this paper will be accepted for publication, I need to apply now for fundings in order to have some chances to obtain any sponsorship to attend the conference. Thus I spent this morning preparing the documentation to request funding from Portuguese Foundation for Science and Technology (FCT) and from Calouste Gulbenkian Foundation.

Tuesday, October 18, 2005

Presentations at 13 EPCG

Last week I've been in Vila Real attending the 13º Encontro Português de Computação Gráfica (13 EPCG). There I presented two short-papers: "Recuperação de Desenhos de Moldes Combinando Imagens e Esboços" and "Simplificação de Plantas de Arquitectura para Classificação Automática". Besides showing our work and get some valuable feedback on it, I saw what other Portuguese researchers on computer graphics are doing.

Listening questions from the audience at 13EPCG
Listening questions from the audience at 13EPCG

Wednesday, October 12, 2005

Submitted paper to WSCG 2006

Uff... It was a tough week. Finally, at the limit of the deadline, we finished and submitted the paper to the 14th International Conferences in Central Europe on
Computer Graphics, Visualization and Computer Vision (WSCG 2006).

Our paper, titled "A Calligraphic Interface for Managing Agents", presents the approach followed during the development of the editION prototype. Unfortunately, due to lack of time, it do not came out as good as we would like to, but we think it was pretty acceptable. Let's wait for the reviewer's comments.

Wednesday, September 28, 2005

QiQA first prototype

The first working prototype of Quick Quotation Assistant (QiQA) are finished. It includes only the basic functionalities in order to be properly tested.

Screenshot of QiQA first prototype

Thursday, September 22, 2005

QiQA interface low-fi prototype

The Quick Quotation Assistant (QiQA) is a tool to support the quotation process in the mouldmaking industry, developed within the scope of the Euro Tooling 21, an Integrated Project (IP) of the EC 6th framework programme. Today we finished the final version of the low fidelity prototype for QiQA's user interface.

QiQA Low-Fi Prototype

Wednesday, September 21, 2005

Updating from VC++6.0 to VC++7.1

I've been using Microsoft Visual Studio 6 for more than five years. Now its time to move on. Therefore, I installed Microsoft Development Environment 2003. Then, I made the necessary changes in some of my projects in order to compile them with MS VC++7.1.

Below I list a simple change that was necessary, but took me lots of time to figure it out:

Where before was:

#include <fstream.h>

Now must be:

#include <fstream>
using namespace std;

So, if you try to compile code in MS VC++7.1 that uses fstream don't forget to specify the namespace you are using. I hope this helps you save precious time.

Tuesday, September 13, 2005

Replacing text in multiple files

Because I changed the GUI framework, I now need to replace in all my source files the string 'wxWindows_2.4.0' by 'wxWidgets-2.6.1'. To that end I devised a small bash script that replaces a string by other in all files of a given path and its subdirectories recurively.

The exec_sed.sh bash script is available here. Feel free to use it and distribute it. If you find any errors or made some improvements, leave a comment here or contact me by e-mail.

Monday, September 12, 2005

Updating from wxWindows to wxWidgets

After a couple of years using wxWindows 2.4.2 on the development of my C++ applications, I decided to upgrade to the latest version of this open-source, cross-platform GUI framework. Due to complains from Microsoft about the wxWindows name, it was changed by August 2004 to wxWidgets. Thus, I now downloaded and compiled wxWidgets 2.61 without major problems.

More information about this library can be found in the wxWidgets project page.

Friday, September 02, 2005

Developing editION

After several weeks without major advances, I'm now working fulltime in the editION prototype. Last week I developed a dummy framework server and implemented a preliminar comunication protocol, while GAIPS team focus on the real framework.

At the same time, some improvements were made to the user interface. For instance, is now possible to change the code associated with each action.

Screenshot of framework dummy server and
editION user interface while editing action details.

Wednesday, August 31, 2005

Summer Holidays

The deserved resting time was over. For three weeks I've been out of office, in my summer holidays with my wife and my son. Now I'm back to work.

Sandra and Tiago at the end of our holidays.

Thursday, July 21, 2005


Yesterday I defended my MSc thesis. Now I am "Master Alfredo", eh eh eh...

Seriously, if you want to view the presentation used for the defense, it is available for download with demo video (ZIP 13,9MB) or without demo video (ZIP 3.8MB).

Presenting my work.

Friday, July 15, 2005

Another Basket Game

Following the success of the first event, Frederico organised another IMMI group basketball game. Twelve members of our group showed up, including Professor Joaquim Jorge. The result doesn't mather, what is important is the fun we all had.

Playing basketball... or, at least, trying to.

Monday, July 11, 2005

Preliminary presentation for MSc thesis defense

I have just finished the preliminary version of the presentation for my MSc thesis defense. A PDF file with the presentation is available here. I will appreciate any comments or suggestions on it.

Wednesday, July 06, 2005

No more CG exams to evaluate!

After three entire days evaluating Computer Graphics exams from a second year course and one morning publishing the results, and I've reached the "end of the tunnel". No more exams to evaluate this semester. Yesss!!!!

Friday, June 24, 2005

The IMMI group basketball game

For the first time, Professor Joaquim Jorge, Frederico Figueiredo and Paula Caetano arranged a baskteball match bewteen IMMI group members, followed by a delicious dinner. It was a funny and relaxing afternoon and evening. Now we are counting the days for the next event.

A picture taken during the game. Everybody is enjoying it.

Wednesday, June 15, 2005

Finished final version of MSc dissertation

Finally I have finished the final version of my dissertation for the degree of Master of Science in Information Systems and Computer Engineering at IST. An electronic version of this document is available for download in PDF (11.8MB) or PS(28.8MB) format.

Friday, June 03, 2005

Started development of editION

I am developing an calligraphic interface for an agents platform. This tool is part of a joint effort between our IMMI group and the GAIPS group to produce a distributed solution to support the development of agent worlds.

Currently, the interface is on the early stage of development. However, we expect that within a few weeks a beta version of this tool will be available.

Screen-shot of an early version of editION.

Thursday, April 07, 2005

New contacts

Because I have moved to Taguspark campus, my postal address and telephone contact had changed. However, mail and home pages addresses remain unaltered. Thus, my current contacts are:

Postal Address:
Intelligent Multimodal Interfaces Group, INESC-ID
Taguspark, Edifício IST, Núcleo 7, Piso 2
Avenida Professor Doutor Cavaco Silva
2780-990 Porto Salvo, Portugal

Tel: +351.21.4233567
Fax: +351.21.4233252

Mail: alfredo.ferreira.jr@inesc-id.pt,
Web page (institutional): http://immi.inesc-id.pt/~afj
Web page (personal): http://afj.planetaclix.pt

Tuesday, April 05, 2005

Moving to Taguspark

Finally, after a long struggle to find a proper room for our group in IST's Taguspark campus, part of the IMMI group are moving there. I was the first to settle down, packing and unpacking my stuff last week. Despite some problems with furniture, which I hope will be solved soon, the room is now usable and all four of us have moved in. Below I depict the actual room layout, while we are waiting for the definitive furniture.

Actual room layout

Wednesday, March 23, 2005

Introduction changed in my Master Thesis Dissertation

During last weeks I have been rewriting my Master Thesis dissertation according to jury comments. One point focused was the "Introduction" chapter, which was a little bit confusing. Thus, it was completely renewed, suffering deep changes. From all changes made, the more important was a novel description for the system architecture of my approach for sketch and image based retrieval of engineering drawings, illustrated below.

SIBR system architecture.

Friday, February 25, 2005

Bentley-Ottmann algorithm implented in PDdemo

Our algorithm for polygon detection (see paper "Polygon Detection from a Set of Lines") starts by detecting the line segment intersections. To that end, we use the Bentley-Ottmann sweep line algorithm [BO79]. Today, we have implemented the intersection reporting functionality in the PDlib library. Thus, the PDdemo application evolved to version 0.2, which is available for download here.

Screenshot of PDdemo in debug mode

[BO79] J.L.Bentley and T.Ottmann. "Algorithms for reporting and counting geometric intersections. IEEE Transactions on Computers, pages 643-647, 1979.

Tuesday, February 01, 2005

Teaching Human-Computer Interfaces is time consuming!

During last weeks all my time was consumed by the Human-Computer Interfaces course I teach in IST. I have been correcting and discussing student's projects. Additionaly, the preparation, correction and revision of exams had took several days. Thus, my PhD is on stand by, as well as the correction of my MsC dissertation thesis.