You are on page 1of 2

Bengali AI CV challenge competition:

Deadline for registration: 30 june 2018


REGISTER:
=============
NO REGISTRATION FEES REQUIRED

Visit the following website, read the rules and request to join the competition:
https://www.kaggle.com/c/numta/

TO LEARN MORE ABOUT THE COMPETITION, JOIN THE COMMUNITY:


https://www.facebook.com/groups/334432463719627/
To contribute or donate visit:
www.bengali.ai

MAXIMUM NUMBER OF PARTICIPANTS IN A TEAM IS 3

BENGALI HANDWRITTEN DIGIT RECOGNITION


========================================
Handwritten digit recognition is a benchmark task in Computer Vision, and has historical importance for artificial
intelligence (AI) research. The Bengali Handwritten Digit Recognition task provides a convenient starting point for
Bengali Optical Character Recognition (OCR) research. We have accumulated a large dataset (85,000+) of Bengali
digits (NumtaDB) which can be used by researchers for benchmarking their algorithm. In this competition, your goal
is to correctly identify digits from six different datasets sourced from different ages, populations and geographic
locations. We have a hidden test set consisting of both natural and augmented pictures of Bengali handwritten
digits. The competition is designed both for beginners and experts.

FOR BEGINNERS:
-------------------------------
- Kaggle kernels (coding platform) have all of the python modules built-in, just login and learn by coding.
- We have starter codes on Kaggle ranging from image processing to deep learning to help you out!!
- Learn from the community, our knowledge-base partners will provide you with support for 24-7. Ask questions, we
will learn while teaching you.

FOR EXPERTS:
-------------------------------
- The test labels are hidden. The test set is prepared in a competitive manner where there are two heavily
augmented subsets. The augmented subsets were rechecked for human recognition where a good number of digits
were hard to recognize, even for humans.
- All the datasets have been partitioned into training and testing sets so that handwriting from the same
subject/contributor is not present in both.

DATASET DESCRIPTION:
=========================
The dataset is a combination of six datasets that were gathered from different sources and at different times.
However, each of them was checked rigorously under the same evaluation criterion so that all digits were at least
legible to one human being without any prior knowledge. Descriptions of these datasets including collection
methodology, image segmentation and extraction and image formats of these datasets are described
in www.bengali.ai/datasets. The sources are labeled from ‘a’ to ‘f’. The training and testing sets have separate
subsets depending on the source of the data (training-a, testing-a, etc.). Dataset-f had no corresponding metadata
for contributors for which all of it was added to the testing set (testing-f). The metric for the competition is selected to
be the Unweighted Average Accuracy (UAA).

PRIZE:
========================
PRIZE TO BE OFFICIALLY ANNOUNCED SOON

DEADLINE:
========================
This is a month long competition which will end in 30th JUNE, 2018
ABOUT BENGALI.AI
=========================
Bengali.AI is a community posing to solve the absence of open sourced datasets for Bengali Natural Language
Processing/Computer Vision research. The datasets open sourced by Bengali.AI go through a rigorous process of
standardization and are licensed under Creative Commons Attribution 4.0 International (CC BY 4.0), open for both
non-commercial and commercial use.

KNOWLEDGE-BASE PARTNER:
===========================
সত্যেন ব োস ব জ্ঞোন ক্লো , ত্ু েট
https://www.facebook.com/satyenbose/

You might also like