Disguised Faces in a Wild Competition

24 views Leave a comment

Image credit: geralt around Pixabay, CC0 Public Domain

With new advancements in low learning, a capabilities of involuntary face approval have been significantly increased. However, face approval in an unrestrained sourroundings with non-cooperative users is still a investigate challenge, impending for users such as law coercion agencies. While several covariates such as pose, expression, illumination, aging, and low fortitude have perceived poignant attention, “disguise” is still deliberate an strenuous covariate of face recognition.

Disguise as a covariate involves both conscious and unintended changes on a face by that one can possibly blear his/her temperament or burlesque someone else’s identity. The problem can be serve exacerbated due to an unrestrained sourroundings or “in a wild” scenarios. However, costume in a furious has not been complicated in a extensive way, essentially due to unavailability of such a database.

As partial of this workshop, we will control a foe in that participants will be asked to uncover their formula on a due sheltered faces in a furious (DFW) database. Top behaving algorithms will be invited to contention their papers in a seminar and comparison papers will be invited for presentation. Authors who have not participated in a foe can also contention a paper.


The foe will have dual endowment winners in any of a following 3 categories:

  1. Impersonation (First place and Runner up)

  2. Obfuscation (First place and Runner up)

  3. Overall correctness (First place and Runner up)

In sum there will be 6 awards with financial prizes as follows:

  • $6,000 USD awarded to a tip scoring acquiescence in any category

  • $2,500 USD awarded to the runner up in any category

Source: iab-rubric.org

Comment this news or article