IMAGE FEATURE EXTRACTION BY DEEP LEARNING MODELS FOR REVERSE IMAGE SEARCH
Article Sidebar
Main Article Content
Abstract
Image feature extraction by deep learning models for reverse image search was aimed to develop a model of the image feature extraction employing deep learning models and evaluate the precise results of reverse image search. This research utilized ResNet50 model, which has been pre-trained transfer learning then well-tuned as the feature extractor for a dataset of 20 Thai-food-image categories that have been internationally popular. The processes mentioned were constructed as a dataset representing a semantic image for comparing with search images using cosine similarity measurement. These processes enabled the fast and accurate image retrieval without the need for labeled data. The research resulted showed the precise evaluation of the reverse image search, especially for the first three results achieving an 80-percent precision. When increasing the number of retrieved results to 5 and 10 images, the precision was at a good level. These results aligned with users’ behavior, who have typically focused only on the top-ranked results. However, similarity among visually alike images, variations in viewpoint, scale, illumination, including background clutter affected errors in recognizing distinctive features. The outcome of this research was an evaluation of the performance of the model employing in the real-situation scenarios, which could serve as a guideline for developing image retrieval systems in e-commerce or duplicating image identification for online media.
Downloads
Article Details
Copyright Notice
The copyright of research articles published in the VRU Research and Development Journal Science and Technology Journal belongs to the Research and Development Institute, Valaya Alongkorn Rajabhat University under the Royal Patronage. Reproduction of the content, in whole or in part, is prohibited without prior written permission from the university.
Responsibility
The content published in the VRU Research and Development Journal Science and Technology Journal is the sole responsibility of the author(s). The journal does not assume responsibility for errors arising from the printing process.
References
เรวัต แสงสุริยงค์. (2565). ความเสี่ยงของการเกิดความคลาดเคลื่อนในการวิจัยเชิงปริมาณด้านสังคมวิทยา. วารสารวิชาการมนุษยศาสตร์และสังคมศาสตร์ มหาวิทยาลัยบูรพา, 30(1), 158-185.
Broz, M. (2024). Photo statistics: How many photos are taken every day?. Retrieved from https://photutorial.com/photos-statistics/
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009). ImageNet: A large-scale hierarchical image database. In 2009 IEEE Conference on Computer Vision and Pattern Recognition (pp. 248–255). Miami, FL, USA: IEEE.
GemmyTheGeek. (2021). Overcome your fear for Thai food with Foody Dudy. Retrieved from https://gemmythegeek.medium.com/overcome-your-fear-for-thai-food-with-foody-dudy-7ff6d7702b22
Gonzalez, R. C., & Woods, R. E. (2008). Digital image processing (3rd ed.). Pearson.
Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep learning. MIT Press.
He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 770–778). Las Vegas, NV, USA: IEEE.
Jansen, B. J., & Spink, A. (2006). How are we searching the World Wide Web? A comparison of nine search engine transaction logs. Information processing &management, 42(1), 248-263
Jolliffe, I. T. (2002). Principal component analysis (2nd ed.). Springer.
Keogh, E., & Mueen, A. (2017). Curse of dimensionality. In C. Sammut & G. I. Webb (Eds.), Encyclopedia of machine learning and data mining (pp. 314–315). Boston: Springer.
Kiros, R., Salakhutdinov, R., & Zemel, R. (2014). Unifying visual-semantic embeddings with multimodal neural language models. arXiv preprint arXiv:1411.2539.
LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436-444.
Manning, C. D., Raghavan, P., & Schütze, H. (2008). Introduction to information retrieval. Cambridge University Press.
Mishra, A., Kumar, A., Yadav, A., Ansari, A., & Gupta, N. (2020). Image Search Engine. International Research Journal of Engineering and Technology (IRJET), 7(5), 2841-2844.
Razavian, A. S., Azizpour, H., Sullivan, J., & Carlsson, S. (2014). CNN features off-the-shelf: An astounding baseline for recognition. In 2014 IEEE conference on computer vision and pattern recognition workshops (pp. 806-813)
Tourism Authority of Thailand. (2021). Eat Like You Wanna Live GASTRONOMY FOR ALL. Tourism Authority of Thailand
Wiens, M. (2017). 40 Thai foods we can’t live without. Retrieved from https://edition.cnn.com/travel/article/bangkok-food-thai-dishes/index.html
