STM Journals Peer-reviewed research from Consortium e-Learning Network

E-Commerce for Future & Trends Original Research

Comparative Study of AI-Driven Fashion Trend Prediction System Using AI and ML: A Review

  1. Shruti Vivekanand Chavan Department of Computer, MET's Institute of Engineering, Bhujbal Knowledge City, College in Nashik
  2. Janhavi Santosh Pardeshi Department of Computer, MET's Institute of Engineering, Bhujbal Knowledge City, College in Nashik
  3. Shubhangi Nivrutti Pingle Department of Computer, MET's Institute of Engineering, Bhujbal Knowledge City, College in Nashik
  4. Raksha Sandip Sonawane Department of Computer, MET's Institute of Engineering, Bhujbal Knowledge City, College in Nashik
  5. Vipin K. Wani Department of Computer, MET's Institute of Engineering, Bhujbal Knowledge City, College in Nashik

Abstract

To overcome the challenges in fashion trend forecasting, researchers have introduced several advanced and data-driven approaches. One such method uses a long short-term memory (LSTM) model combined with an encoder-decoder architecture to extract meaningful fashion content and recognize styles from product images. This model achieves higher accuracy in predicting upcoming fashion trends by incorporating varying price intervals and has shown impressive results when evaluated on the Amazon fashion dataset. Another forecasting model focuses on analyzing the evolution of fashion styles over time. It can detect and predict new combinations of styles, identify recurring patterns, and distinguish between emerging trends and timeless classics. This model's effectiveness has been validated using a dataset of 80,000 fashion products sold on Amazon over a period of 6 years. In addition, the Fashion Attributes Recognition Network (FARNet) has been developed to enhance attribute prediction by simultaneously identifying multiple fashion attributes and correcting noisy labels, which commonly occur in large-scale datasets. FARNet has demonstrated substantial improvements over previous methods and has been applied to the RichWear dataset, comprising over 322,000 images sourced from an Asian social media platform. Through image clustering and refined label prediction, FARNet has proven useful in identifying regional and street fashion trends, especially within Asian markets. Together, these models represent a significant advancement in the ability to predict, analyze, and adapt to changing fashion landscapes using deep learning and large-scale visual data.

Keywords

References (10)

  1. Al-Halah Z, Stiefelhagen R, Grauman K. Fashion forward: forecasting visual style in fashion. In: Proceedings of the IEEE International Conference on Computer Vision (ICCV), Venice, Italy, October 22–29, 2017. pp. 388–397.
  2. Jin B, Lu B, Wu H, Shi W, Li Y. Fashion style forecasts based on different price ranges. In: Proceedings of the 3rd International Conference on Consumer Electronics, Communications and Networks (CECNet), Chongqing, China, March 12–14, 2013. pp. 509–512. doi:10.1109/CECNet.2013.6703534.
  3. Huang FH, Lu HM, Hsu YW. From street photos to fashion trends: leveraging user-provided noisy labels for fashion understanding. IEEE Access. 2020; 8: 209671–209682. doi:10.1109/ACCESS.2020.3038566.
  4. Grabe I, Zhu J, Agirrezabal M. Fashion style generation: evolutionary search with Gaussian mixture models in the latent space. In: Proceedings of the 11th International Conference on Computational Creativity (ICCC), Coimbra, Portugal, September 7–11, 2020. pp. 307–314.
  5. Gu C, Xie H, Lu X, Zhang C. CGMVAE: coupling GMM prior and GMM estimator for unsupervised clustering and disentanglement. IEEE Access. 2021; 9: 63794–63807. doi:10.1109/ACCESS.2021.3074229.
  6. Shi M, Chussid C, Yang P, Jia M, Dyk Lewis V, Cao W. The exploration of artificial intelligence application in fashion trend forecasting. Textile Research Journal. 2021;91(19-20):2357-2386. doi:10.1177/00405175211006212
  7. Wani V, Baghe M, Gupta H. A comparative study of image fusion techniques based on feature using transforms function. Int J Emerg Technol Adv Eng. 2013; 3 (11): 438–443.
  8. Thoiba Singh N, Chhikara L, Raj P, Kumar S. Fashion Forecasting using Machine Learning Techniques. 2023 IEEE International Conference on Integrated Circuits and Communication Systems (ICICACS). 2023:01-06. doi:10.1109/icicacs57338.2023.10099844
  9. Speight L, Karpova E. Analysis of Fashion Trend Forecasting Industry: Challenges, Opportunities, and Outlooks. Innovate to Elevate. 2022. doi:10.31274/itaa.15872
  10. Chang AA, Cynthia, Devita, Ramadhan JF, Adnan ZKS, Kanigoro B, et al. Fashion Trend Forecasting Using Machine Learning Techniques: A Review. Lecture Notes in Networks and Systems. 2021:34-44. doi:10.1007/978-3-030-90321-3_5
Support