[Gallery] Generative images for dance

I am a super fun of dance and I’m always fascinated by the limits of body expression and how creative choreographers are when I happen to see a good piece. One of the reasons why I love dance so much is that the story that comes through the entire body,  from the beginning to the end. I think dance is so unique that most of the current main media won’t be able to show it’s meaning/characteristics.  So I started to dig through some dance pieces and want to bring them into another art forms.

So here I’m presenting my experiments with different methods of media integration about dance.

The journey began with a happy accident. Last year I went to Kate Ware’s performance “Unstruck”. “Unstruck”explores in movement and sound a process of live formation. Like stones in a river, or sound through a tunnel. And it is about energy, physical exchange.  It was shown in conjunction with another exhibition Clifford Ross: “Water | Waves | Wood”.  I felt like there was another force of power when these two things happen together: the picture and the movement.   

  

Therefore I started thinking about generative new media for dance that can show the feature of dance itself.

  1. Style transfer based on deep learning:

These works are implementations of “A Neural Algorithm of Artistic Style” by L. Gatys, A. Ecker, and M. Bethge in August 2015. Style transfer is a method for transferring the artistic style of one input image onto another using convolutional neural network. I experimented this method with some pieces of dance I like.

  • Unstruck by Kate Weare: (keyword: wave, energy)

unstruck0

Original content image from the screenshot of  “Unstruck” ‘s documentation video:

4

The style image from the exhibition Clifford Ross: Water | Waves | Wood: 

A more dynamic gif version to better represent the dance:

unstruck_deeplearning

 

Other style transfers:

  • Folding by Shenwei dance arts: (keyword: fold)

sample_4sample_1 sample_2 sample_3  sample_5

Original content image and style image:

  b

 

  • Folding by Shenwei dance arts: (keyword: fold)

folding

Original content image and style image:

  a

 

  • Folding by Shenwei dance arts: (keyword: fold)

folding_2_green folding_2_red

Original content image and style image:

  b

 

  • Folding by Shenwei dance arts: (keyword: fold)

62eb26fab26cdb6336220f060beec5b1

Original content image and style image:

  qaSpHGoWhuiIFyldxOrYXNLCR

 

  • Map by Shenwei dance arts: (keyword: pattern, rotation, bouncing etc.)

mao

Original content image and style image:

AwPuzYXBGSxlRZspebfWKEvOD  hDoFHrXiMcGnQztWJLmySwTBd

 

  • Map by Shenwei dance arts: (keyword: pattern, rotation, bouncing etc.)

map

Original content image and style image:

AwPuzYXBGSxlRZspebfWKEvOD  diOIcyVSpBAlwHKTRtaFhqvNm

 

  • Map by Shenwei dance arts: (keyword: pattern, rotation, bouncing etc.)

4091848d9fef4301d2c31f39a9c92509

Original content image and style image:  kryUoRIWQeqXgiSjJpMZwxPma

 

0923c73818f4dc0de6a7d3268130e3bc

Original content image and style image:

8083158279_775bdf793c_b  aps_timespace

 

5f59fe6a878d696c167453b3b931d344

Original content image and style image:

  

 

 

2. Photo Mosaic based on keywords image search 

In the field of photographic imaging, a photographic mosaic, also known under the term Photomosaic, a portmanteau of photo and mosaic, is a picture (usually a photograph) that has been divided into (usually equal sized) tiled sections, each of which is replaced with another photograph that matches the target photo.When viewed at low magnifications, the individual pixels appear as the primary image, while close examination reveals that the image is in fact made up of many hundreds or thousands of smaller images. Most of the time they are a computer-created type of montage.(from Wikipedia)

 

I integrated the mosaic algorithm here https://github.com/cinemast/OpenMosaic with google image custom search.  The keywords are crawled from the different websites about each piece, currently I’m not able to tell the accuracy about these keywords(if they truly reflect the feature of the dance).

  • Folding by Shenwei dance arts: (keywords: flesh, fold, shenwei, fabric, paper etc.)

folding_1 folding_3folding_0

 

Keyword: Cultural specificity, deviously beautiful, giggly young women, jagged path, drug addiction, intervention

Number of division: 21

Original :

Keyword: phenomenal cover, Drag king, punky garage-rockers Spider, cultural specificity

Number of division: 17

  • Gif version of Tool is Loot, playing with different resolution and keywords:

 

Keywords: unnecessarily heavy eye, disingenuously anti-classical way, problematic new works, special guest stars

Number of devides: 47

47__The Bright Stream__0__ unnecessarily heavy eye__disingenuously anti-classical way__problematic new works__special guest stars

Original :

Keywords:male Season fairies,tense dramatic action,sublime ballerina type

Number of devides: 42

 

There are many more mosaics you can see here: http://restage.streamgao.com/wsearch/loadmore?loadmore=See+previous+examples.  They are made collectively by the users of my thesis restage.

My thesis is about meta data/meta media crawling for dance and generation of the photo mosaic. You are definitely welcome to see this post:  http://www.streamgao.com/?p=2285    and Try the demo here: http://restage.streamgao.com/