0% found this document useful (0 votes)
69 views13 pages

Final PPT

Uploaded by

raihan262006
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
69 views13 pages

Final PPT

Uploaded by

raihan262006
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd

"AI Image Generator"

Submitted in fulfillment of the requirements for the award of


Diploma in Computer Engineering
SUBMITTED TO

MAHARASHTRA STATE BOARD OF TECHNICAL EDUCATION, MUMBAI

PRESENTED BY

Name of Student(s) Enrollment No


1. Shaikh Sania Sarwar 2211650048
2. Shaikh Mohammed Raihan Iqbal 2211650077
3. Faizan Jatu 2211650069

Project Guide :Prof.Ruksar Attar

Diploma in Computer Engineering


JMCT Polytechnic, Nashik
Introduction
The ability to generate images from text is a rapidly advancing field in
artificial intelligence, with the potential to revolutionize creative
industries and empower everyday users. By harnessing the power of
deep learning and natural language processing, AI-powered image
generation can transform textual descriptions into visually stunning and
conceptually rich representations, opening up new avenues for artistic
expression, product design, and visual storytelling
Literature Review
Paper Title Authors Year Description

A Comprehensive Shaw, Kashyap, Sahil, 2023 Comprehensive


Review on Generative Dwivedi, Khandelwal, review of generative
AI - Text To Image Sharma AI text-to-image
Generator models, covering their
techniques and
applications.

Generation of Images Yadav, Sinha, Jain, 2024 Exploration of AI-based


from Text Using AI Agrawal, Francis techniques for generating
images from text
descriptions, focusing on
their strengths and
limitations.
Problem Definition
Current image generation models often struggle to represent uncommon
entities accurately. This can result in inaccurate images, which is
problematic for real-world applications. The challenge lies in bridging
the gap between textual descriptions and accurate visual representations,
especially when dealing with rare or unseen entities.
Proposed Methodology
1. User Interaction & Input
• Users register and log in to the web platform.
• Once authenticated, they enter a text prompt describing the image they wish to
generate.
2. Request Processing (Flask Backend)
• The backend (Python + Flask) receives the user prompt.
• Flask handles session management, routes, and API requests securely.
3. AI Model Invocation
• The system sends the prompt to a text-to-image API (e.g., Hugging Face's
implementation of Stable Diffusion).
• The model interprets the prompt and generates one or more images.
4. Image Response & Display
• The generated image(s) are returned as a response.
• Flask passes the images to the frontend for rendering.
5. Storage and History Management
• Generated images, along with the corresponding prompts and timestamps, are
stored in a MySQL database.
• This enables personalized image history and retrieval for each user.
6. Frontend Rendering
• The frontend (HTML, CSS, JS) displays the output in a clean, responsive interface.
• Users can view, download, or regenerate images.
Flow Diagram
User Interface
1

Text Prompt Submission


2

Flask Backend Server (Request)


3

AI Model API (Stable Diffusion via Hugging Face)


4

Flask Backend (Stores)


5

Output
6
Requirements

Hardware Software

High-performance GPUs are necessary for the Suitable software tools are required for development,
computationally intensive tasks of training and generating coding, and model training.
images.
• Flask
• Intel processor i5 or higher • Python 3.6
• 8 GB RAM • HTML/CSS/JavaScript
• 500 GB hard disk • MySQL

• Hugging Face API / Stable


Diffusion
Advantages
Unleash Creativity
Automate the image creation process, enabling users to explore and experiment with
ideas.

Improved Productivity
Generate custom visuals on demand, enhancing workflows in design, visualization, and
e-commerce.

Increased Accessibility
Democratize visual content creation, making it accessible to individuals with limited
artistic skills.

Fostering Innovation
Rapid exploration and iteration of visual ideas can drive innovation in various
industries.
Disadvantages

Lack of Human Touch


AI-generated images may lack the creativity and delicate qualities of human-made visuals.

Limited Customization
AI image generators may not meet the specific needs of every organization, particularly
those with unique branding requirements.

Dependency on Data
Inaccurate or incomplete data can lead to substandard image generation, emphasizing
the importance of quality data input.

Ethical Concerns
Future Scope
This project aims to contribute to advancements in text-to-image
generation, with potential applications in various fields. The future holds
opportunities for improvement in image quality, diversity of generated
content, and the ability to handle more complex and abstract
descriptions. Furthermore, research can focus on addressing ethical
concerns and developing responsible guidelines for the use of AI image
generation.
Conclusion
In conclusion, this text-to-image generation
project presents a compelling opportunity to
unlock new creative possibilities and enhance
productivity across various industries. By
leveraging the power of AI, users can
effortlessly generate custom visuals to support
their needs, from design to e-commerce and
beyond.
References
The project proposal references two key studies
on text-to-image generation using AI:

1. 1.A Comprehensive Review on Generative


AI- Text To Image Generator By JETIR.
2. 2.Generation of Images from Text Using AI
By MECS Press.

You might also like