Advertising design images - Extracting some factors

This project is about NLP (natural language processing).

PLEASE JUST BID IF YOU KNOW NLP.

We need to process two factors in some advertising design images.

We have processed about 16000 images and extract this two factors.

For each image these data is available:

1. Text:

A list of strings with a font size factor like this:

[{text: "sample text": size: 10}, {text: "sample text 2": size: 4}]

10 or 4 is a percent value calculated from text size compared to image size. For example:

(text height in pixel) / (image height in pixel) * 100

2. Color

All colors in the image pixel by pixel is indexed to 16 main colors and then the area size that has been taken by each color is calculated and finally we have a percent value for each color:

{0: 20, 1: 4, 2: 1, ..., 15: 0}

AND WHAT WE NEED:

We need a freelancer familiar with NLP to process this data and extract meaningful and significant relations between text and colors in ad. images.

For example we can calculate rate of each color for a specific word, but this is very simple.

What we need is to finding complex relations with consideration of NLP knowledge, like role of each word in the sentence and ...

We have plan to extract more factors from our sample repository.

So this project will CONTINUE and this is the first one.

Just bid if you have NLP knowledge and experience.

You SHOULD SEND EVIDENCE to show that.

Grafik Tasarımı Doğal Dil Photoshop

Proje NO: #4699364

Proje hakkında