Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{177567,
        author = {Aditi Bandaru and Harshini Pasupuleti and Bhavana maneesha and B. Arun and Mr. Bhaskar Das},
        title = {RECOMMENDATION OF FASHION CLOTHS USING MACHINE LEARNING MODEL},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {11},
        number = {12},
        pages = {1340-1344},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=177567},
        abstract = {Color compatibility is the essence of fashion and dress selection, not only determining the beauty of an outfit but also determining consumer purchasing behavior. Following the significance of visually pleasing and harmonious garments, this study develops a deep learning-based system for color categorization and fashion suggestion automatically. Its main goal is to intelligently pair shirts with similar bottomwear based on established color theory principles to improve user experience in fashion use cases like virtual try-ons, personal shopping assistants, and e-commerce.
Its core is a Convolutional Neural Network (CNN) architecture built with TensorFlow that can classify shirts into pre-defined color classes from images. These are basic colors like Red, Black, White, Yellow, and Pink. The CNN model is trained and tested on specially preprocessed data, split into labeled directories, stored in ZIP archive format. Data includes high-resolution images of bottomwear and shirts colored. In order to have better generalization and stability, the images are preprocessed through the use of the Keras ImageDataGenerator class that performs a sequence of augmentation operations such as rotation, scaling, shearing, and flipping in addition to normalization.
The model, having been trained, is very efficient in multi-class classification and can correctly classify the color of the input shirt image. Classification leads to a rule-based color pair engine producing esthetically balanced bottomwear recommendations. The engine relies on a database of traditional and modern color harmony rules of color theory—complementary, analogous, triadic, and split-complementary relations—to recommend colors for bottomwear with greater aesthetic balance and contrast.
For the effective retrieval and output of suggested products, there exists a tidy CSV metadata file within the process. A file of this type will have one record per article of clothing with the correct attributes including article type, color tag, and image file path. Pandas library is utilized to read this metadata, and filtering the relevant bottomwear option according to color rules of compatibility is also done by it. Lastly, the chosen recommendation is shown graphically to the user using OpenCV for image processing and Matplotlib for layout visualization, thus adding to a collaborative visual and user-friendly system of recommendation.
This solution leverages both the capabilities of deep learning and rule-based reasoning to provide a scalable and intelligent personalized fashion recommendation solution. Future research can investigate extending the application of user preference, garment type, seasonal fashion trend, and texture analysis to further improve the quality and flexibility of recommendations across different categories of fashion.},
        keywords = {},
        month = {May},
        }