Skip to main content

RESOURCES FOR MEMBERS

CENPARMI offers members access to labs, databases, user guides and other tools that facilitate collaboration and consultation.

Databases


CENPARMI members continually make training databases available to researchers and students:

  • CENPARMI Handwritten Digits Database
  • Arabic Cheques Database
  • English/French Cheques Database
  • Arabic Isolated Words Database (69 words)
  • Arabic Isolated Words Database (40 words)
  • Arabic Isolated Digits, Letters, Special Symbols, Numeral Strings, Dates Database

This list is always growing. Please contact Nicola Nobile at nicola@cenparmi.concordia.ca for any questions about CENPARMI databases and resources.


The CENPARMI Human Reticulocyte Dataset

The CENPARMI human reticulocyte dataset is an image dataset for classification problems. This dataset consists of 2461 instances that belong to 3 classes:

  1. Reticulocytes,
  2. Erythrocytes,
  3. and Unknown background objects.

In order to obtain the dataset, the following consent form must be completed: https://forms.gle/fFYFvgebaQotooDQ6

More information can be found on the dataset wiki: https://github.com/Rabiah86/Animal2Human/wiki

Samples:

0011.png 0019.png 0020.png
0031.png 10_0020.png 10_C0365.jpg
15_C0428.jpg   1_C0197.jpg
C0399C.jpg   C0403.jpg

Counts:

count.png

HeroSvm


HeroSvm is a high-performance library that improves training SVM (Support Vector Machine) classification on large training sets. HeroSvm was developed by former PhD student Jianxiong Dong.

Local section (for members only)


The local section has additional information about labs, databases and general guidelines for use. The local section is restricted to CENPARMI members only.

Back to top Back to top

© Concordia University