maps & mulch

prostheticknowledge:

Dwitter

Social Network by Andreas Løve Selvik for Javascript computer graphics coding presents visual experiments in 140 characters or fewer, and all run slick in-browser.

See for yourself here

Oh; apparently I’ve been busy elsewhere in my years off tumblr.

(Source: dwitter.net)

Comments (View) and 279 notes since 3 years ago / View (Permalink) →

prostheticknowledge:

Neural Image Analogies and Neural Doodles

Computer Science research from Adam Wentz and Alex Champandard (coincidentally, not together) extends neural network style transfer method allowing to dictate how image sections should be stylisticly composed and rendered (in two different ways).

The initial concept was put together by MRL NYU in 2001, and here is a demostration video to highlight the concept:

We present a new framework for processing images by example, called “image analogies.”  Rather than attempting to program individual filters by hand, we attempt to automatically learn filters from training data.

… our system “learns” the transformation from A to A’, and then applies that transformation to B to get B’.  In other words, we compute B’ to complete the analogy.

More Here

Then Adam Wentz has updated the work to faciliate contemporary developments and technology:

This is basically an implementation of this “Image Analogies” paper http://www.mrl.nyu.edu/projects/image-analogies/index.html In our case, we use feature maps from VGG16. The patch matching and blending is done with a method described in “Combining Markov Random Fields and Convolutional Neural Networks for Image Synthesis” http://arxiv.org/abs/1601.04589  Effects similar to that paper can be achieved by turning off the analogy loss (or leave it on!) and turning on the B/B’ content weighting via the --b-content-w parameter.

The Abe Lincoln image above was put together by Audun Mathias Øygard using Adam’s work.

Code can be found here

Alex Champandard has also developed his own code and paper in this field which he calls ‘Neural Doodles’:

Convolutional neural networks (CNNs) have proven highly effective at image synthesis and style transfer. For most users, however, using them as tools can be a challenging task due to their unpredictable behavior that goes against common intuitions. This paper introduces a novel concept to augment such generative architectures with semantic annotations, either by manually authoring pixel labels or using existing solutions for semantic segmentation. The result is a content-aware generative algorithm that offers meaningful control over the outcome.  Thus, we increase the quality of images generated by avoiding common glitches, make the results look significantly more plausible, and extend the functional range of these algorithms—whether for portraits or landscapes, etc.  Applications include semantic style transfer and turning doodles with few colors into masterful paintings!

His research paper can be found here

Comments (View) and 1,022 notes since 7 years ago / View (Permalink) →

prostheticknowledge:

Protopiper

Fabrication prototyping tech from the HPI Human Computer Interaction Lab is a computerized tapegun that can produce 3D forms to help visualize objects in spaces:

Protopiper is a computer aided, hand-held fabrication device that allows users to sketch room-sized objects at actual scale. The key idea behind protopiper is that it forms adhesive tape into tubes as its main building material, rather than extruded plastic or photopolymer lines. Since the resulting tubes are hollow they offer excellent strength-to-weight ratio, thus scale well to large structures. 

More Here and Here

(Source: hpi.de, via prostheticknowledge)

Comments (View) and 1,612 notes since 7 years ago / View (Permalink) →

http://cdn1.tnwcdn.com/wp-content/blogs.dir/1/files/2015/09/fn63697_03.jpg

(via This hilarious Cisco fail is a network engineer’s worst nightmare)

(Source: thenextweb.com)

Comments (View) and 0 notes since 7 years ago / View (Permalink) →