ArtVise AI Innovation Brief
The SIP brief for ArtVise AI outlines a student-led innovation project focused on developing a secure local application for artists and designers. Since it runs directly on the user’s computer, the application can function both online and offline while ensuring privacy through local data processing. It integrates AI technology to provide personalized feedback on artwork while addressing data security concerns. The project aims to enhance creative growth by offering constructive critiques and references to similar artists, fostering a safe and supportive environment for artistic development and learning.
AUTHOR: Sanyerlis "Vanessa" Camacaro
Technical Field
This project involves the technical fields of Artificial Intelligence.
Background Information
From exhibiting AI artwork across more than 20 countries to pursuing a degree in AI engineering, my journey as a Digital AI artist has been fueled by a passion for blending photography with prompt engineering. However, amidst my immersion in the art community, I observed a prevalent issue: artists' reluctance to embrace AI due to privacy and security worries. This insight inspired me to create ArtVise AI, a secure application that runs locally on the user’s computer, allowing it to work both offline and online while providing artists and designers with safe, constructive feedback on their work and safeguarding their intellectual property. Additionally, ArtVise AI functions as a locally running AI chatbot that answers general questions without compromising users' privacy. By addressing these concerns, I aim to foster trust in AI technology, empowering artists to explore and innovate freely.
Prior Art
ArtVise AI incorporates concepts inspired by existing innovations in AI chatbot, image analysis and description generation. The following prior art has influenced my project:
2. Midjourney – Art Generator with Option to Describe Image
Midjourney offers tools for generating and describing artistic images using AI.
3. Pally - Image Description Generator
Pally provides AI-powered image description generation capabilities.
4. Ahrefs - Free AI Image Alt Text Generator
Description: Ahrefs offers a tool for generating alternative text for images using AI.
5. KichangKim/DeepDanbooru – Image Description Generator
DeepDanbooru, by KichangKim, is an AI model for generating image descriptions.
6. Photor.io - Analyze your Photos
Photor.io provides AI-based photo analysis capabilities.
7. Fly.io - Open Source AI for Image Description
Fly.io offers open-source AI models for generating image descriptions.
ArtVise AI distinguishes itself from these applications by being a local application that provides offline capabilities for image analysis. While some, like ChatGPT and Midjourney, require membership for use, ArtVise AI runs directly on the user’s computer, offering both offline and online functionality without recurring costs. Moreover, in addition to image description and analysis, ArtVise AI provides image rating and style referencing, enhancing its utility for artists and designers seeking comprehensive feedback and improvement suggestions. Additionally, ArtVise AI operates as a locally running AI chatbot that answers general questions while protecting users' privacy, further broadening its appeal and usefulness beyond just the art community.
Project Description
This student innovation project aims to create ArtVise AI, an offline AI chatbot designed to provide artists and designers with constructive feedback on their artwork without the need to upload images online. The primary goal is to address privacy and security concerns while facilitating the growth and improvement of artists through AI-powered critiques. Additionally, ArtVise AI functions as an offline AI chatbot that answers general questions while protecting users' privacy. ArtVise AI targets artists and designers who value privacy and seek high-quality, reliable feedback to enhance their creative process. The tool is intended for individual artists, art schools, and artistic communities, providing a secure environment for artistic growth and collaboration.
Innovation Claim
ArtVise AI revolutionizes the art critique process by providing a secure application on the user’s computer, allowing them to receive detailed feedback on their artwork without risking privacy. Since it runs locally, ArtVise AI can work both offline and online while ensuring full control of user data. It also functions as a private AI chatbot that answers general questions without compromising security. This innovative approach makes it easier and safer for users to improve their craft and seek assistance, promoting widespread adoption among those concerned about online data security.
Usage Scenario
ArtVise AI, originally designed as a secure local application for providing feedback on creative work, can expand its utility across a multitude of disciplines. Running directly on the user’s computer, it can function both offline and online. For instance, an architect could utilize ArtVise AI to enhance building design by receiving architectural insights, design principles, and blueprint feedback. Similarly, a student studying abroad could rely on ArtVise AI for language learning, improving pronunciation and grammar through personalized feedback. In the culinary realm, both home chefs and professionals can benefit from the AI's recipe suggestions and cooking tips, refining their culinary skills. Athletes and fitness enthusiasts can personalize workout plans and receive nutritional advice from ArtVise AI to optimize their fitness routines. Individuals seeking stress management and mindfulness practices can find support through ArtVise AI, which offers strategies for emotional well-being. In addition, event planners can streamline large-scale event organization with marketing, ideas, and logistics support from ArtVise AI. Lastly, enthusiasts learning new skills such as photography or coding can accelerate their progress with educational resources and practical tips offered by the AI, demonstrating its versatility across diverse fields.
Evaluation Criteria
1. Does ArtVise AI function without requiring an internet connection?
2. Can ArtVise AI be executed solely through the command line interface?
3. Can users upload images in specified formats (PNG, JPEG, GIF, BMP) up to 12MB without encountering issues?
4. Does the AI provide constructive and relevant feedback on uploaded artwork?
5. Is the feedback provided by ArtVise AI comparable in quality to that from human critics?
6. Are privacy and security concerns adequately addressed by the offline nature of ArtVise AI?
7. Is the user interface intuitive and easy to navigate?
8. Can the AI accurately reference three artists with similar styles based on uploaded artwork?
9. Does ArtVise AI support side-by-side image display for comparisons?
10. Are all File and Edit functionalities (New conversation, save conversation, load conversation, cut, clear, copy and paste) fully operational?
11. Are all core functionalities (image analysis, rating, and feedback) fully operational?
12. Can ArtVise AI provide responses to general questions besides art-related queries?
13. Are there no critical bugs or performance issues observed during testing?
14. Can ArtVise AI handle requests from users outside of art-related queries, such as architecture, language learning, culinary advice, fitness tips, stress management, event planning, photography, and coding?
15. Is ArtVise AI's performance consistent across various usage scenarios, maintaining quality feedback and usability?
16. Is the documentation comprehensive, clear and accessible?
17. Does the final demonstration of ArtVise AI meet all specified project requirements effectively?
18. Can ArtVise AI provide users with real-time metrics, such as the total runtime of the application and the processing time for each individual user input, allowing users to track performance and responsiveness?
Objectives and Tasks Associated with the Project
Objective 1: Develop Core Functionalities
Objective 1: Develop Core Functionalities
Objective 1: Develop Core Functionalities
- Task: Complete integration of Llama and Llava15ChatHandler models (Expected Completion Date: 06/20/2024).
- Task: Develop image uploading feature (Expected Completion Date: 06/20/2024).
- Task: Implement text-based interactions (Expected Completion Date: 06/20/2024).
- Task: Create initial AI response mechanism (Expected Completion Date: 06/20/2024).
Objective 2: User Interface Development
Objective 1: Develop Core Functionalities
Objective 1: Develop Core Functionalities
- Task: Design and implement GUI using Tkinter (Expected Completion Date: 06/22/2024).
- Task: Integrate side-by-side image display feature (Expected Completion Date: 06/26/2024).
- Enhancing Overall Appearance (Expected Completion Date: 11/04/2024).
Objective 3: Testing and Debugging
Objective 1: Develop Core Functionalities
Objective 4: Finalization and Showcase Preparation
- Task: Conduct initial rounds of testing (Expected Completion Date: 07/30/2024).
- Task: Enhancing AI model to respond faster (Expected Completion Date: 11/04/2024).
- Task: Perform thorough QA (Expected Completion Date: 11/04/2024).
- Task: Address identified issues (Expected Completion Date: 11/04/2024).
- Task: Refine user interface interactions (Expected Completion Date: 11/04/2024).
Objective 4: Finalization and Showcase Preparation
Objective 4: Finalization and Showcase Preparation
Objective 4: Finalization and Showcase Preparation
- Task: Finalize documentation (Expected Completion Date: 01/15/2025).
- Task: Creating Portfolio SIP Website (Expected Completion Date: 07/15/2024).
- Task: Prepare for SIP showcase (Expected Completion Date: 01/15/2025).
- Task: Ensure all project requirements are met (Expected Completion Date: 01/15/2025).
Completed:
Objective 4: Finalization and Showcase Preparation
Completed:
1. Completed integration of Llama and Llava15ChatHandler models. (Completed 06/20/2024).
2. Developed image uploading feature (Completed 06/20/2024).
3. Implement text-based interactions (Completed 06/20/2024).
4. Created initial AI response mechanism (Completed 06/20/2024).
5. Designed and implemented GUI using Tkinter (Completed 06/22/2024).
6. Integrated side-by-side image display feature (Completed 06/26/2024).
7. Created logo (Completed 06/29/2024).
8. Gantt Chart (Completed 07/02/2024).
9. First User Testing (Completed 07/04/2024).
10. Scheduling initial rounds of testing (Completed 07/30/2024).
11. Creating Portfolio SIP Website (Expected Completion Date: 07/15/2024).
12. Art Evaluation Metric added to code. (Completed 09/01/2025).
13. Updated SIP Brief Document.
(Completed 09/19/2025).
14. Second User Testing.
(Completed 09/19/2025).
15. Added real-time metrics, such as the total runtime of the application and the processing time for each individual user input.
(Completed 09/19/2025).
In Progress:
1. Enhancing AI prompt personality to be more consistent with answers.
(Expected Completion Date: 09/26/2025).
2. Add Installation instructions on Github READ Me file.
(Expected Completion Date: 09/26/2025).
3. Update ArtVise AI Simple Gantt Chart with completed tasks.
(Expected Completion Date: 09/26/2025).
4. Update Prior Art with the latest applications in the market today.
(Expected Completion Date: 09/26/2025).
5. Schedule an appoinment with my SME to get feedback on my current code.
(Expected Completion Date: 09/26/2025).
Final Task:
1. Conduct final demonstration of ArtVise AI (Expected Completion Date: 12/19/2025).
Description of Design Prototype
ArtVise AI is being developed as a local application and offline AI chatbot on a desktop platform using Python, Tkinter, and integrated with the Llama and Llava15ChatHandler models. Running directly on the user’s computer, it can function both offline and online, ensuring that all functionalities can be executed without relying on an internet connection, addressing privacy and security concerns for users. The prototype includes several core features and a user-friendly interface:
Core Functionalities:
- Image Analysis and Rating: Users can upload images in formats such as PNG, JPEG, GIF, and BMP (up to 12MB). The AI analyzes the artwork, providing constructive feedback and ratings.
- Style Referencing: The AI references three artists with similar styles based on the uploaded artwork, aiding artists in finding inspiration and understanding their artistic influences.
- General Queries: Beyond art-related feedback, the AI chatbot can answer general questions, enhancing its utility.
- ArtVise AI provide users with real-time metrics, such as the total runtime of the application and the processing time for each individual user input, allowing users to track performance and responsiveness?
User Interface:
- GUI Design: Built with Tkinter, the interface includes areas for image display, text interactions, and side-by-side image comparison.
- Functionality: The interface supports all core features, including image uploading, feedback display, response timer and file management (new conversation, save, load, cut, clear, copy, and paste).
Development and Execution:
- Code Writing: The code for ArtVise AI is written using Visual Studio Code, providing a robust development environment with extensive support for debugging and extensions.
- Running the Prototype: To run ArtVise AI, you need to locate the folder containing the project on your computer, open the command line through the folder window, and type ‘python ArtviseAI.py’. This will initiate the GUI to interact with ArtVise AI. The system requires a Python environment with Tkinter and the integrated AI models. The offline nature ensures no data is uploaded or shared, maintaining user privacy.
Artwork Evaluation Rating Scale:
1 - Very Poor
Quality: Extremely low resolution, highly pixelated.
Clarity: Indistinguishable elements; very difficult to discern any details.
Composition: Lacks structure and organization.
Completeness: Major elements missing; appears unfinished.
Color/Lighting: Poorly managed; overly dark or bright.
2 - Poor
Quality: Low resolution, noticeable blurriness.
Clarity: Difficult to distinguish elements.
Composition: Disorganized; lacks cohesive arrangement.
Completeness: Several key components missing.
Color/Lighting: Inconsistent and unbalanced.
3 - Below Average
Quality: Below average resolution; some blurriness.
Clarity: Key details are somewhat obscured.
Composition: Basic attempt at organization, but awkward.
Completeness: Important details missing; partially incomplete.
Color/Lighting: Uneven; lacks harmony.
4 - Fair
Quality: Fair resolution; slight blurriness.
Clarity: Generally clear, but some details are not well-defined.
Composition: Basic structure; elements somewhat awkwardly placed.
Completeness: Mostly complete; minor details missing.
Color/Lighting: Basic use; lacks vibrancy.
5 - Average
Quality: Average resolution; minor sharpness issues.
Clarity: Clear enough to see most details.
Composition: Adequate organization of elements.
Completeness: Generally complete; most details present.
Color/Lighting: Acceptable use; neither remarkable nor poor.
6 - Above Average
Quality: Good resolution; generally clear.
Clarity: Clear with minor ambiguities.
Composition: Thoughtful arrangement of elements.
Completeness: Complete with most details present.
Color/Lighting: Effective use; adds depth.
7 - Good
Quality: High resolution; sharp and clear.
Clarity: Most details are easily distinguishable.
Composition: Strong and cohesive arrangement.
Completeness: Fully detailed.
Color/Lighting: Good use; enhances overall impact.
8 - Very Good
Quality: Very high resolution; crisp and clear.
Clarity: Extremely clear; all details easily seen.
Composition: Well-balanced and harmonious.
Completeness: Very detailed and complete.
Color/Lighting: Excellent use; significantly enhances artwork.
9 - Excellent
Quality: Exceptional resolution; near-perfect clarity.
Clarity: Crystal clear; every detail sharp.
Composition: Highly engaging and balanced.
Completeness: Fully detailed with exceptional precision.
Color/Lighting: Masterful use; significantly enhances artwork.
10 - Masterpiece
Quality: Outstanding resolution; flawless clarity.
Clarity: Perfect clarity; every minute detail visible.
Composition: Masterful arrangement; meticulously placed.
Completeness: Entirely complete; no missing elements.
Color/Lighting: Perfect use; elevates artwork to the highest level.
Evaluation Plan
The evaluation plan for ArtVise AI involves comprehensive testing to ensure it meets all specified criteria, ensuring functionality, performance, and user experience are up to standard:
- Offline Functionality: Verify that ArtVise AI operates entirely offline without any need for an internet connection.
- Command Line Execution: Ensure the prototype can be executed and operated from the command line, including all functionalities.
- Image Handling: Test the image uploading feature with various formats (PNG, JPEG, GIF, BMP) and sizes (up to 12MB) to confirm seamless operation.
- Feedback Quality: Compare AI-generated feedback with human critiques to evaluate the relevance and constructiveness of the suggestions provided by ArtVise AI.
- Privacy and Security: Assess the offline capabilities to confirm that no data is transmitted, ensuring user privacy is maintained.
- User Interface: Evaluate the intuitiveness and navigability of the GUI, ensuring it is user-friendly and efficient.
- Style Referencing: Test the accuracy of the AI in referencing three artists with similar styles based on the uploaded artwork.
- Response Timer: Ensure that response timers are active as soon as the application starts. After the AI chatbot provides a response, it should display both the total duration the application has been running and the time the AI took to respond to the user’s input.
- File and Edit Functionalities: Verify the functionality of file operations (new conversation, save, load, cut, clear, copy, and paste).
- Core Functionalities: Test all core features (image analysis, rating, style referencing) for performance and accuracy.
- General Queries: Assess the AI's ability to respond to non-art-related questions effectively.
- Performance and Bug Testing: Conduct thorough testing to identify and resolve any critical bugs or performance issues.
- Versatility: Test the AI's performance across various usage scenarios, including architecture, language learning, culinary advice, fitness tips, stress management, event planning, photography, and coding.
- Documentation: Ensure the documentation is comprehensive, clear, and accessible, providing users with all necessary information.
- Final Demonstration: Prepare for a final demonstration to showcase ArtVise AI's capabilities, ensuring all project requirements are met.
Appendices
Appendix A: ArtVise AI Logo – avatar.PNG
Appendix B: ArtVise AI Simple Gantt Chart – Sanyerlis_Camacaro_ArtVise_AI_Simple_Gantt_chart.xlsx
Appendix C: SIP Pitch - Camacaro_Sanyerlis_SIP311_Pitch.pptx
Appendix D: Trello Board – TrelloSanyerlis.jpeg
Appendix E: ArtVise AI Prototype - ArtViseAIPrototype.jpg
GITHUB PROJECT CODE
This website uses cookies.
We use cookies to analyze website traffic and optimize your website experience. By accepting our use of cookies, your data will be aggregated with all other user data.