Evotegra: Extend 100k promo dataset with license plates and faces (adapted to R11)

Name of Project: Extend 100k promo dataset with license plates and faces

Proposal in one sentence

Extend the round 8 promotional 100k traffic dataset with annotations for licence plates and faces to enable the creation and evaluation of automated anonymization algorithms.

Description of the project and what problem is it solving:

In Ocean DAO Round 8 we commited to create a professional traffic image dataset with at least 100.000 images in ~300 categories and an estimated million objects. The dataset will soon be published on the Ocean Market with an initial data token price of 1 Ocean. The idea is to create promotional datasets that help to establish Ocean Market as a prime source of high quality data.
We propose to extend this promotional dataset next the promised 300 classes with annotations for faces and license plates. By adding these 2 classes we add significant extra value to the dataset which then can be used to train and evaluate automated anonymization algorithms which is an important requirement to assure compliance with privacy regulation such as GDPR.

Category

Unleash data

Metric

Data Consume Volume

What is the final product?

The final product is an extension of the promotional image dataset proposed in DAO round 8 with 2 additional classes “license plates” and “faces”.

Grant Deliverables:

  1. Extend the promotional 100k dataset with face annotations
  2. Extend the promotional 100k dataset with license plate annotations
  3. Provide 5000 annotated images

R8 Deliverables

Our image collection, annotation, and publishing process follows a waterfall-like structure.
Although we have 2.5% of the R8 grant carried over, we are realizing the expenses for this batch of data now (Round 8), and amortizing the payment/work over the next 1-2 months.

The following tasks will be carried over from round 8:
[ ] Publish the dataset in a pool for the initial price of 1 Ocean (~1 month). Publishing will take place after annotation is completed. This will cost a final 2.5% of the grant available.

Funding Requested:

3k USD

ROI

We see this effort as a starting point to add other promotional datasets. Our goal is to establish Ocean as a prime source for high quality data. The attractive entry price should encourage many first time users to get familiar with wallets, Ocean tokens and the Ocean Market itself.
Due to the funding restrictions for the unleash data category in R11 we have to limit the number of annotations we can add to the dataset to 5000. By enabling new use cases to train and test automated anonymization algorithms this will notably increase the value of the data.

Proposal Wallet Address:

0x61B15998893cC746B46C08FEdEE13a0d1b33bBa9

Team

Web: www.evotegra.de
Email: manthey@evotegra.de
Twitter: @evotegra
Country of Residence: Germany

Previous OceanDAO Grant?

Yes

1 Like

Of course I will support you and I am biased.

1 Like

A larger batch of images can be annotated more efficiently. Therefore we propose to carry over the 5k images to R12.

We finished the annotation of 5000 images. The results will be published along along with the dataset with OceanV4.