PROJECT RUNWAY ML
By tairan hao
Project Brief: Exploring the different package programs in runway ML, and creating varies output from the data gathered from municipal bond art museum. Also introducing one program/algorithm models. I chose cnOCR - which is a AI system that can recognize Chinese character in different kinds of images.
Tool: Runway ML
Duration: 1 week
Credits: Runway ML models publisher
CNOCR MODEL INTRODUCTION
It is a model that can recognize as much as Chinese words, letters in a picture.(Chinese letters only)
The Factor that will affect the output result -
Background color
2. The complexity of words combinations
3. Contrast between the background of color and the text color
4. The size, density of the image
5. The setting (the simple/multiple)
the choice Simple indicates the simple line of words, if putting complex words as input, it may not recognize the words perfectly.
The choice multiple indicates the complex line of words.
INPUT: IMAGES
OUTPUT: Coding lines(all words recognized in the input image)
MODEL EXPLORATIONS
MODEL: SPACE COCO
MODEL: ADAPTIVE STYLE TRANSFER
MODEL: DENSE POSE
ARTS INPUT & OUTPUT
INPUT 1:
CHIARO/SCURO SERIES, NO.1, 2017
Watercolor on Arches 140 lb. watercolor paper
16 x 20 in
OUTPUT 1:
from BIGBIGAN model
INPUT 2:
"CAN YOU LOVE TWO PEOPLE AT THE SAME TIME?", 2019
Digital C-print / Unique hand cut collage
8.5 x 10.5 in
OUTPUT 2:
from BIGBIGAN model
OUTPUT 3:
from BIGBIGAN model
INPUT 4:
CIRCLES OF ORANGES, BERRIES, GRAPES, 2017
Monoprint with Akua Intaglio ink on paper
12 x 9 in
OUTPUT 4:
from BIGBIGAN model