I have implemented some thresholding methods and need some standard database for testing. Can someone suggest some sites for classification of image object and background? It will be better if the database has reference image, which can be useful to check the performance of thresholding methods.
With thresholding do you mean image segmentation? For some evaluation dataset I raccomend you to start from some well known papers in the field, that I've linked below.
http://groups.inf.ed.ac.uk/calvin/Publications/kuettel12cvpr.pdf
http://www.robots.ox.ac.uk/~vilem/NIPS2011.pdf
Hi,
you can use the stanford bake of VOC dataset, or the VOC segmentation dataset...just group all objects a foreground.
http://dags.stanford.edu/projects/scenedataset.html
http://pascallin.ecs.soton.ac.uk/challenges/VOC/
These will quite challenging for a threshold.....
Cheers
Paul
Check section "Experiment Dataset and Ground Truth" for link to Berkeley Segmentation Dataset & Segmentation evaluation tool
http://kspace.cdvp.dcu.ie/public/segmentation-evaluation/downloads.html
Hi. There are three publicly available data sets of retinal fundus images that come with the ground truth of the blood vessels tree. They are commonly used to test segmentation algorithms.
Here the links:
DRIVE
http://www.isi.uu.nl/Research/Databases/DRIVE/
STARE
http://www.parl.clemson.edu/~ahoover/stare/probing/index.html
CHASE_DB1
http://blogs.kingston.ac.uk/retinal/chasedb1/
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Processing
- Image Processing