Computer Vision for Creative Optimisation: KPI maximisation through image analysis

Sample Dynamic creative optimisation

Project details

Table of contents

• Introduction
• Overview
• Objective
• Data
• Requirements
• Install
• Using the application
• Screenshots
• Notebooks

Introduction

This week challenge mainly focus on the Computer vision and deep learning creative advertised,on the Adluio company. The assignment is to use computer vision techniques based on deep learning to creatively optimize mobile advertising. Computer vision technology has transformed the world by allowing machines to achieve human-level understanding of images and videos. The success of deep learning based computer vision has led to a number of novel applications such as Autonomous driving for cars, tumour detection from X-rays, farm weed detection and yield prediction from satellite and drone images, visual try-on features for clothes and jewellery in e-commerce, and many more.

Objective

• Adludio, an online mobile advertising company, is our customer this week. The following service is offered to customers by Adludio. Create a creative, often known as an interactive advertisement. A rich advertisement including interactive features such as a mini-game engine, video, text, and photos is referred to as a creative. On behalf of a client, makes these creatives available to viewers. Real-time bidding is used by Adludio to purchase impressions on an open market. Utilizing sophisticated machine learning algorithms, optimizes the process of creative designing while also focusing on guaranteed inventory.
• Adludio has been running a huge number of commercials, and each one has a unique Creative. These designs were produced based on the designers' prior work and the requirements of the business. There is no way to assess creatives during production and predict how well they might perform once they are served. Adludio aims to address this issue by creating an algorithm that enables it to optimize its creatives in light of campaign performance information. We are entrusted with creating a computer vision system based on deep learning that separates items from creative assets and associates them with KPI criteria of the related campaigns in order to achieve that.

Data

The Dataset Archive is composed of an ‘Assets’ folder and a ‘performance_data’ CSV file. As the names suggest, the folder contains the asset image used to build the creatives and the CSV file contains the performance values for each creative. Each folder within the ‘Assets’ folder is a ‘game_id’ string value, and its reference performance is available in the CSV file.

• Game_id - Represents a unique identifier of a creative for the performance values.
• ER - Represents the engagement rate score of the creative.
• CTR - Represents the click-through rate score of the creative.

Requirements

• Python > 3.5
• Docker
• pip install -r requirements.txt

Installation

Use the package manager pip to install OpenCV.

pip install opencv-python

Screenshots

The detailed use and implementation

Notebooks

All the notebooks that are used in this project including EDA, data cleaning and summarization along with some machine learning model generations are found here in the Notebooks folder.

Scripts

All the scripts and modules used for this project relating to interactions with kafka, airflow, and other frameworks along with default parameters and values used will be found here, in the scripts folder.

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

👤 Birhanu Gebisa

• Email, GitHub, LinkedIn

👤 Birtukan Kumma

• Email, GitHub, LinkedIn

👤 Fisseha Estifanos

• Email, GitHub, LinkedIn, Twitter

👤 Kibatu Woldemariam

• Email, GitHub, LinkedIn

👤 Margaret Chepkirui

• Email, GitHub, LinkedIn

👤 Nahom Habtemichael

• Email, GitHub, LinkedIn

👤 Yohanes Gutema

• Email, GitHub, LinkedIn

Show us your support

Give us a ⭐ if you like this project, and also feel free to contact us at any moment.