Recognition of blurred faces using Local Phase Quantization (2008) at  http://ieeexplore.ieee.org/search/wrapper.jsp?arnumber=4761847

* In this paper, recognition of blurred faces using the recently introduced Local Phase Quantization (LPQ) operator is proposed. LPQ is based on quantizing the Fourier transform phase in local neighborhoods. The phase can be shown to be a blur invariant property under certain commonly fulfilled conditions. In face image analysis, histograms of LPQ labels computed within local regions are used as a face descriptor similarly to the widely used Local Binary Pattern (LBP) methodology for face image description. The experimental results on CMU PIE and FRGC 1.0.4 datasets show that the LPQ descriptor is highly tolerant to blur but still very descriptive outperforming LBP both with blurred and sharp images.


Blurred face recognition via a hybrid network architecture (2000) at  http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?arnumber=906198

* We introduce a hybrid recognition/reconstruction architecture that is suitable for recognition of images degraded by various forms of blur. This architecture includes an ensemble of feedforward networks each of which is constrained to reconstruct the inputs in addition to performing classification. The strength of the constraints is controlled by a regularization parameter. Networks are trained on original as well as Gaussian-blurred images, so as to achieve higher robustness to different blur operators. Face recognition is used to demonstrate the proposed method and results are compared to those of classical unconstrained feedforward architectures. In addition, the effect of state-of-the-art restoration methods is demonstrated and it is shown that image restoration with the proposed hybrid architecture leads to the best and most robust results under various forms of blur
*


When you hide a face by the means of a blur filter then your image processing software simply utilises an algorithm (a specific sequence of steps of calculations) in order to modify some pixels by replacing their colour with the colour of another pixel.

The image processing experts have now found a way to recognise faces from blurred photos, inverting (in practice) the sequence of the algorithm (by performing the same steps and calculations in the reverse order). This explanation is not fully correct in the scientific sense, but it is the only way to explain it to you in simple terms without resorting to complex mathematics (I study machine vision doing an MRes in the South East, and while working on my dissertation I freaked out after I stumbled upon these articles!) I just hope you can appreciate the danger!!

What this means is that when you upload a photo with blurred faces, the cops can simply process your photo with a forensics program and reveal your faces without the blurring effect! It's as simple as 1-2-3 or A-B-C!

The good news though is that the solution is even simpler: you just refrain from using the blurring filter and you paint the whole face with your image processing software's pen by only using one simple colour (black or white), but NEVER the skin's colour because if you copy it then it MAY prove useful for analysis! see this:

Contribution of color to face recognition (2002) at  http://web.mit.edu/bcs/sinha/papers/yip_sinha_ColorFaces.pdf

* One of the key challenges in face perception lies in determining how different facial
* attributes contribute to judgments of identity. In this study, we focus on the role of color cues.
* Although color appears to be a salient attribute of faces, past research has suggested that it
* confers little recognition advantage for identifying people. Here we report experimental results
* suggesting that color cues do play a role in face recognition and their contribution becomes
* evident when shape cues are degraded. Under such conditions, recognition performance with
* color images is significantly better than that with gray-scale images. Our experimental results
* also indicate that the contribution of color may lie not so much in providing diagnostic cues to
* identity as in aiding low-level image-analysis processes such as segmentation.
*



By replacing all the face's pixels with another colour all the information is lost for good (except if you use an image format that keeps undo levels or other metadata) and no one can recover it (not even their all-powerful god).

But I have more for you... there is also a program that can recognise the same person in two different photos as long as she or he wears the same pair of jeans!! have a look at this:

A Garment in the Dock; or, How the FBI Illuminated the Prehistory of A Pair of Denim Jeans (1998-2004) at  http://mcu.sagepub.com/cgi/content/abstract/9/3/293

* This article looks at research carried out at the FBI Laboratory’s Special Photographic Unit in the identification of denim trousers from bank surveillance film. This research, which was published in 1998, showed that despite the ubiquity of jeans, each pair has individual identifying characteristics caused by the manufacturing process and by wear, and that these might be used as evidence in the identification of criminal suspects. What the FBI research also inadvertently illuminated was an otherwise hidden relationship between garment, maker and wearer, in an effective - if accidental - reversal of commodity fetishism.


Enlighten your comrades, translate it, and remember: knowledge is power! Freedom to everyone!