In the context of temporal image forensics, it is not evident that a neural network, trained on images from different time-slots (classes), exploits solely image age related features. Usually, images taken in close temporal proximity (e.g., belonging to the same age class) share some common content properties. Such content bias can be exploited by a neural network. In this work, a novel approach is proposed that evaluates the influence of image content. This approach is verified using synthetic images (where content bias can be ruled out) with an age signal embedded. Based on the proposed approach, it is shown that a deep learning approach proposed in the context of age classification is most likely highly dependent on the image content. As a possible countermeasure, two different models from the field of image steganalysis, along with three different preprocessing techniques to increase the signal-to-noise ratio (age signal to image content), are evaluated using the proposed method.

In ultrasound imaging the appearance of homogeneous regions of tissue is subject to speckle, which for certain applications can make the detection of tissue irregularities difficult. To cope with this, it is common practice to apply speckle reduction filters to the images. Most conventional filtering techniques are fairly hand-crafted and often need to be finely tuned to the present hardware, imaging scheme and application. Learning based techniques on the other hand suffer from the need for a target image for training (in case of fully supervised techniques) or require narrow, complex physics-based models of the speckle appearance that might not apply in all cases. With this work we propose a deep-learning based method for speckle removal without these limitations. To enable this, we make use of realistic ultrasound simulation techniques that allow for instantiation of several independent speckle realizations that represent the exact same tissue, thus allowing for the application of image reconstruction techniques that work with pairs of differently corrupted data. Compared to two other state-of-the-art approaches (non-local means and the Optimized Bayesian non-local means filter) our method performs favorably in qualitative comparisons and quantitative evaluation, despite being trained on simulations alone, and is several orders of magnitude faster.

Bayesian image processing [1] based on a probabilistic graphical model has a long and rich history [2]. In Bayesian image processing, one constructs a posterior distribution and then infers restored images based on the posterior distribution. The posterior distribution is derived from a prior distribution that captures the statistical properties of the images. One of the major challenges of Bayesian image processing is the construction of an effective prior for the images. For this purpose, a Gaussian Markov random field (GMRF) model (or Gaussian graphical model) is a possible choice.

It is a glimpse of life in average - taken from thousands on internet snaps. Researchers have created software that can trawled the internet for snaps - and combine them to create an average or everything from a wedding kiss to a car. They say the software could dramatically improve facial recognition and online shopping - and show that in fact, most wedding snaps are almost identical. Researchers analysed hundreds of Google search results to create these average images of a computer, sunset and a yacht. Life in average: These images, created by artist Jason Savalon, combine 100 images each to produce an'average' image.