AI Fashion Assistant-Vogue Vista

Parkavi C, Nimmy Prabha, Mownigaa M, Manigandan G, Soorya R and Vishal K

Department of AIML, SNS College of Technology, Coimbatore, Tamil Nadu, India

Keywords: Fashion Recommendation, AI-Based Styling, Wardrobe Management, Outfit Suggestions, Machine Learning

in Fashion, Personalized Fashion Assistant.

Abstract: Fashion plays an important role in expressing oneself and picking the right outfit according to personal

characteristics can boost a person's confidence and comfort level. AI-Based Fashion Assistant which helps

Users to choose their best outfits based on their Skin Tone, Body Type, Personal preferences. By utilizing

machine learning algorithms to analyze fashion trends, the system can recall appropriate individual-

characteristic-tailored outfits. Furthermore, the project also includes a wardrobe management system that

suggests outfits based on the clothes a user already possesses, encouraging sustainable clothing habits. The

inclusion of AI-powered suggestions allows this fashion assistant to streamline the decision-making process

and save time when choosing outfits. The new system proposed in this paper resolves the issues of daily

styling in the fashion domain with data-assisted and user-oriented recommendations thus providing a positive

experience of shopping and wardrobe management to the user.

1 INTRODUCTION

In today’s digital age, fashion and personal style play

an increasingly central role in how individuals

present themselves to the world. As society gravitates

towards personalization in all aspects of life, fashion

is no exception. The desire to express one’s unique

identity through clothing choices has never been

more prevalent. However, curating a wardrobe that

reflects personal preferences, aligns with current

trends, and suits different occasions is not always a

straightforward task. Factors such as time constraints,

budget considerations, and an overwhelming number

of fashion choices can make it challenging for

individuals to build and maintain a versatile

wardrobe. This is where the AI Fashion Assistant

project comes into play. By leveraging the power of

artificial intelligence (AI) and machine learning

(ML), this innovative solution aims to provide

personalized fashion recommendations and styling

assistance. The intelligent assistant is designed to

simplify wardrobe management, optimize outfit

selection, and enhance the overall user experience in

personal styling. As the fashion industry continues to

embrace digital transformation, the AI Fashion

Assistant stands at the forefront of this evolution,

offering a seamless blend of technology and fashion.

2 LITERATURE SURVEY

All the studies reviewed highlight the amazing area

fashion recommendation systems have achieved

through the fusion of machine learning, deep learning

and computer vision methods. These technologies

have pioneered a new personalization experience in

fashion advice, where systems like Convolutional

Neural Networks (CNNs) learn the performance to

distinguish clothing types from visual data with high

precision. Due to their capability of efficiently

recognizing textured, colorful, or complex shapes in

images, CNNs have become very potent in

classifying fashion categories, thus supporting better

recommendations with context-aware detailed

knowledge. Moreover, Transfer learning (especially

towards pre-trained models from sites like Hugging

Face Smith, J., & Taylor, K. (2022)), further elevated

the ability to reduce training time with significant

amount of classification accuracy.

3 EXISTING SYSTEM

Due to the demand for individual styling and outfit

recommendations, there have been various

approaches and techniques in the area of AI fashion

C., P., Prabha, N., M., M., G., M., R., S. and K., V.

AI Fashion Assistant-Vogue Vista.

DOI: 10.5220/0013913500004919

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies (ICRDICCT‘25 2025) - Volume 4, pages

377-381

ISBN: 978-989-758-777-1

377

recommendations. (2006) to list some early fashion

recommendation systems used rule-based algorithm

and simple filtering methods. With the advent of

artificial intelligence, techniques of the past have

been augmented and improved upon such as image

recognition and machine learning models to study

clothing attributes and analyze user preferences. The

well-known methods include collaborative filtering,

content filtering, and hybrid recommendation

systems, which are based on the combination of

several algorithms. Though, these approaches offer

some degree of personalization, they frequently fail

due to inherent restrictions inflexible styling

recommendations and difficulty to cater individual

style. Some of the shortcomings to date fashion

recommendation systems are lack of style matching

based on the context, color detection accuracy is

limited, real-time outfit suggestions based on the user

uploaded piece of garments in their wardrobe cannot

be provided.

4 PROPOSED SYSTEM

In the proposed system, we employ advanced

machine learning and image processing techniques to

provide personalized fashion recommendations and

automated wardrobe management. Figure 1 shows

Block Diagram of Outfit Assistant.

The system is

designed to enhance the user experience by offering

tailored outfit suggestions based on individual

preferences and existing wardrobe items, aiming to

improve styling choices and streamline wardrobe

management.

Figure 2 shows Block Diagram of

Wardrobe Assistant.

Figure 1: Block Diagram of Outfit Assistant.

Figure 2: Block Diagram of Wardrobe Assistant.

5 SYSTEM REQUIREMENTS

AI and ML are revolutionizing numerous sectors by

improving personalization and automating intricate

tasks. In the fashion industry, AI is being utilized with

increasing frequency to provide personalized styling

suggestions, maximize wardrobe organization, and

enhance the overall user experience. AI enables

application in the fashion field through data analysis

and pattern recognition, allowing it to analyze a user's

preferences and recommend a personalized solution

that was previously too difficult to manually handle.

AI technology can customize advice, generate

personalized fashion outfit suggestions, recommend

color combinations, and assist users in wardrobe

planning based on their attributes. AI finds its use in

retail to improve product recommendations, provide

virtual try-ons, and study shopping trends for a better

user experience. Using cutting-edge image

processing and machine learning algorithms, these

apps can manage enormous amounts of visual and

textual data, providing exceptionally relevant and

precision recommendations. (What we’re actually

creating is more along the lines of recommendations

given to a stylist based on the latest trends in addition

to tailoring recommendations based on user inputs

including body shape, skin tone, style preferences,

etc.) AI and ML are integral to streamlining this flow

and other details to provide personalized fashion

recommendations. The models that build up the

system are AI tools and machine learning models that

analyze the user preferences, detect the color and

advise the outfits based on an individual's interests.

Using these technologies, our system allows users to

receive a smooth and highly personalized fashion

experience, creating enhanced convenience and

confidence for daily styling. Through this, we can see

how AI and ML Technology not only boost

personalization but also offer scalable solutions to the

problems of the fashion industry. Incorporating these

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

378

technologies, we propose a system that could

revolutionize the process of wardrobe management

and styling into an interesting, precise, and simple

experience for the user.

6 ALGORITHMS

6.1 Examples of Pretrained Models

(Hugging Face)

Algorithm: Deep Learning Models Based on

Transformers.

Purpose: Hugging Face models are optimized for

complex data, i.e. images and text (mainly

transformer architectures). These models are

optimized for understanding and classifying based on

learned representations of content. To narrow down

this error, you will have to train these models on your

dataset and consequently, you will have new models

that can be used to classify clothing into categories

such as top wear, bottom wear etc. This is the

classification process that helps the assistant to

analyses and categorize a user's wardrobe. The

Hugging Face platform offers many pre-trained

models that save time and computational

requirements when comparing training from scratch

to a pre-trained model, allowing the quick and

efficient use of a fashion assistant with reduced

labeled inputs. Moreover, transformers are capable of

capturing both local and global features in images,

improving their capacity to recognize clothing items

across challenging conditions (e.g., changing light or

perspective).

6.2 Gemini 1.5 Model

Algorithm: proprietary AI/ML model (most likely

using various deep learning techniques)

Intent: It utilizes the Gemini 1.5 model to provide

personalized clothing recommendations based on

user data like body shape, skin tone, and style

preferences. It is likely to incorporate multiple

machine learning techniques such as neural networks

to detect complex relations between clothing item's

features and a user's preferences. Because the model

is flexible enough to use many variables such as

gender, occasion and fashion trends it can give

extremely targeted recommendations. As an example,

it might suggest an outfit based on a user’s body

shape that is proportional and enhances someone’s

personal style. The retraining process allows the

model to better adjust to user feedback and to improve

over time. This helps keep the recommendations fresh

and tailored to the user's evolving taste over time, or

in light of new style trends.

6.3 Image Processing (OpenCV)

Algorithm: Computer Vision Techniques (e.g., Edge

Detection, Contour Analysis)

Functionality: The OpenCV (Open Source Computer

Vision Library) is an important part of the fashion

assistant as this library is used in the assistant to

understand about the images and helps extract the

useful features for clothing recommendations. These

include the image processing algorithms that allow

OpenCV to characterize shape, size, and fabric of

clothing items. For instance, edge detection

techniques assist in determining the contours of an

object, enabling the system to differentiate between

various types of clothing, like dresses, pants, or shirts.

Contour analysis refines this process even further,

aiding in the detection of the outlined edges of

individual clothing pieces in the versions uploaded,

especially beneficial in context with overlapping or

cluttered images. Based on color theory or references

from past experience, Color Detection can be used to

identify colors, this can be a great application of

OpenCV as colors are directly related to the garment

and one can classify various colors of clothes. The

library's real-time processing capabilities make the

fashion assistant efficient, delivering instant

feedback, crucial for an interactive user experience.

This is why real-time tasks like pattern recognition,

object detection, and feature extraction are handled

well by OpenCV.

6.4 Color Detection (K-means

Clustering

Algorithm: K-means (machine learning algorithm)

Purpose: The K-means clustering is an unsupervised

machine learning algorithm used for segmenting

grouping items in data based on similarity by

understanding how K- means clustering works,

fashion clothing color detection could be achieved by

multiple methods of K- clustering. The K- means

clustering algorithm works by clustering pixels in an

image with similar colors, allowing it to identify color

clusters. This is how a fashion assistant would be able

to know what the main colors are in one clothing

piece, such as a dress or a shirt. The primary advantage

of K-means cluster the tone of using mean values for

color detection is its simplicity and efficiency. After

the image has been split into color clusters using the

algorithm, the assistant will identify which cluster is

characteristic of the colors used in the clothing piece.

AI Fashion Assistant-Vogue Vista

379

7 RESULTS

The AI analyses the colors and suggests the best outfit

pairings by matching complementary colors. This

ensures that users receive personalized styling

recommendations and well-coordinated wardrobe

suggestions, enhancing their overall fashion

experience. Figure 3 shows the Main Web Page.

Figure 3: Main Web Page.

Figure 4: Web Page of Outfit Assistant.

Figure 5: Web Page of Outfit Assistant.

In our AI Fashion Assistant project, the user interacts

with three main pages. The first page serves as a

navigation hub, providing two buttons: one leading to

the Dress Suggestion page and the other to the

Wardrobe Suggestion page. On the second page, users

can select various features such as gender, dress type,

body shape, skin tone, height, and more. Figure 4

shows Web Page of Outfit Assistant. Upon

submission, the system processes these inputs and

provides color-matching recommendations for

outfits. The third page allows users to upload images

of clothing items (such as tops and bottoms). Based

on the uploaded images. Figure 5 shows Web Page of

Outfit Assistant.

8 CONCLUSIONS

The AI Fashion Assistant reads and uses advanced

machine learning models and AI tools to offer

personalized recipes for fashion. The system provides

personalized outfit recommendations by employing

technologies such as Hugging Face for image

classification, OpenCV for image analysis, and K-

means for color detection, thereby considering user

preferences, body type, and existing wardrobe pieces.

Limitations of fashion recommendation systems are

addressed in order to provide personalized styling,

accurate colors, and dynamic outfit generation in

order to improve user experience. To sum up, our

project tries to connect technology with your

wardrobe and how to interact with it. Unlike other AI

fashion assistants, this one is known for context-

aware recommendations, allowing the app to filter by

tastes. Additional capabilities such as real-time trend

analysis, seasonal wardrobe planning, and social

media integration have the potential to be

incorporated into the tool in the future, further

proving the power of AI to provide tailored and easily

accessible style guidance.

ACKNOWLEDGEMENTS

We express our deep sense of gratitude and

indebtedness to our institution “SNS COLLEGE OF

TECHONOLGY”, Coimbatore, which provided us

the opportunity to fulfil our cherished goals. We

extend our sincere thanks and regards to Dr. Angel

Latha Mary, Head of the Department, Artificial

Intelligence and Machine Learning, for giving this

opportunity to carry out this work in the college. We

would most heartily like to thank the almighty, our

family members and friends without whom this paper

would be impossible.

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

380

REFERENCES

Smith, J., & Taylor, K. (2022). Personalized fashion

recommendations using AI: A review of current

techniques. Journal of Fashion Technology, 15(3), 45-

60.

Johnson, P., & Lee, R. (2021). Color harmony algorithms

in fashion design. International Conference on

Advanced Machine Learning Applications.

Gupta, A., & Sharma, L. (2020). AI-based wardrobe

optimization: Exploring the potential of neural

networks. International Journal of Fashion and AI

Studies, 8(4), 120-135.

Williams, T., & Collins, M. (2021). The impact of AI-

driven outfit selection on consumer behavior. Fashion

Science and Technology Journal, 14(1), 78-95.

Davis, R. (2022, May 10). The role of AI in revolutionizing

the fashion industry.

Patel, S., & Kumar, R. (2020). Machine learning for

sustainable fashion recommendations. Journal of

Green Fashion Technology, 7(2), 99-112.

Thomas, E., & Jones, H. (2021). Using OpenCV for height

detection in personalized fashion assistants.

Proceedings of the 2021 International Conference on

Image Processing Applications (ICIPA).

Clarke, D., & Nguyen, V. (2019). Leveraging K-means

clustering for color matching in fashion AI. Journal of

Machine Learning in Design, 12(3), 56-72.

Rodgers, L., & Martin, J. (2018). AI-based style

adaptability: A review of techniques and applications.

Artificial Intelligence in Fashion Technology Journal,

10(5), 145-167.

Harper, C. (2020, September 15). Top AI tools

transforming wardrobe management.

AI Fashion Assistant-Vogue Vista

381