Cette page appartient aux archives web de l'EPFL et n'est plus tenue à jour.
This page belongs to EPFL's web archive and is no longer updated.

USwisTrack

Report 20080613
The enlosed file gives a description of work carried out this week and the results of the work at this moment.

Regards,
Harsha
Posted by Harsha Umesh Babu at 17:27
Report 20080609
Last 2 weeks of work has been summarized here and the future outline is discussed as known at the time of writing this report.

> Based on discussions with Greg, Pierre 2 weeks before, the goals were set to identify the Squares and possible Dimers/Trimers.

> The Squares are now being detected to good extent, rigorous testing is planned for next week to ascertain the efficiency of the algorithm.

> Dimers are also being detected and and being stored as array of Squares; after discussion with Pierre, its better that a connectivity matrix on the lines of Edge matrix be made for Squares so that not only Dimers, more larger combinations can also be easy made out.

Next weeks work
> Rigor Testing of the Squares detection algorithm on generated images.
> Rigor testing on real images
> Explore the possibility of temporal evaluation of the images.
Posted by Harsha Umesh Babu at 9:47
Report 20080523
This is the weekly summary. Theres an doc attached which pictorically summarizes my work. Hope to have discussion with Pierre and Greg on Monday to see how we can proceed from here.
Posted by Harsha Umesh Babu at 14:14
Report 20080520
The summary of work carried out till now is attached as a word document.

Regards,
Harsha
Posted by Harsha Umesh Babu at 11:11
Report Week 3
This weeks activities are summarized below.

Literature Study
> Continued Image registration but gave up quickly after realizing theres not much matching and tracking thats involved.
> Tried to understand harris corners, more with demos than theory.

Research - in line with discussion with Greg
> trying to get all edges of an image to be clearly represented. Mainly playing around with imageJ and trying out few things.
> Blob approach, getting to fill out cubes and treat them as blobs and see if I can get any metrics out of them. Blobs are formed....analyzing them and getting metrics is added to next weeks activity.
> Detecting corners, tried to implement and iterative algorithm but found that harris method already exists and a very good implementation of it by Peter Kovesi. Trying to see if I can use it and get any info out f images.

Activities planned for next week
> Define some form of metrics and try and extract them from the processed images.


Harsha U.B.
Posted by Harsha Umesh Babu at 16:43