Loading article...
Loading article...
Stream/Area:
-
Subject:
-
Paper ID:
IJSREM68886
Volume:
10 |
11 |
Nov 2026
Pages:
1-6
Publication Frequency:
Monthly (12 issues per year)
License:International Journal of Scientific Research in Engineering and Management follows the Creative Commons Attribution 4.0 (CC BY 4.0) License, allowing free use and distribution with proper attribution.
Department of Computer Science and Engineering
Don Bosco Institute of Technology
Department of Computer Science and Engineering
Don Bosco Institute of Technology
Department of Computer Science and Engineering
Don Bosco Institute of Technology
Department of Computer Science and Engineering
Don Bosco Institute of Technology
Department of Computer Science and Engineering
Don Bosco Institute of Technology
Online shopping normally assumes that the customer can name what they want. That assumption breaks down when a product is unfamiliar, or when a single photograph shows several products at once. This paper presents a web application that takes an image, typed text, or spoken audio and turns whatever it recognises into product searches on Amazon and eBay. Four YOLO-based detection models find the objects in an image, and a custom fusion step, weighted by confidence and based on intersection over union (IoU), merges overlapping boxes of the same class into one. A convolutional neural network (CNN) trained for 20 epochs on selected categories of the Stanford Online Products dataset reaches a reported classification accuracy of 89%. An optical character recognition (OCR) module reads printed text in the image, such as brand names or model numbers, so the search queries can be more specific. The system is written in Python with the Flask framework and an HTML/CSS front end, and a single page shows the uploaded image, the annotated image, the detected objects, the extracted text and the retrieved product listings. The work suggests that joining detection, classification, text reading and marketplace search in one pipeline can make product identification easier for non-expert users.
KEYWORDS
This paper presented an AI-driven application that identifies multiple objects in user images and links them to product listings. The system combines four YOLO-based detectors with a custom confidence-weighted IoU fusion step, a CNN classifier with a reported accuracy of 89% on selected Stanford Online Products categories, an OCR module, and Amazon and eBay search integration, all delivered through a Flask web interface. By letting users start from a picture instead of a product name, the system makes product identification and online product discovery simpler. Future work includes a quantitative evaluation of the fusion step against the individual models, extending the CNN to more product categories, applying image–text embeddings to retrieval [18], ranking listings by visual similarity, and supporting additional shopping platforms.
[1] W. Diwu, R. Zhang, L. Feng, and Q. Mu, “Image recognition and product recommendation algorithms based on deep learning,” in Proc. 3rd Int. Conf. Distrib. Comput. Electr. Circuits Electron. (ICDCECE), 2024, doi: 10.1109/ICDCECE60827.2024.10549427. [2] S. Bhoir and S. Patil, “Transfer learning with deep neural networks for image classification in the e-commerce industry,” in Proc. IEEE 7th Int. Conf. Converg. Technol. (I2CT), 2022, doi: 10.1109/I2CT54291.2022.9824903. [3] C. Zhang and J. Feng, “Cross-modal product image retrieval for e-commerce recommendation systems via deep learning,” Traitement du Signal, vol. 42, no. 4, pp. 2289–2299, Aug. 2025, doi: 10.18280/ts.420438. [4] L. Liu, Y. Wang, and W. Chi, “Image recognition technology based on machine learning,” IEEE Access, 2020, doi: 10.1109/ACCESS.2020.3021590. [5] A. M., G. Parasa, et al., “Image recognition & classification with machine learning,” in Proc. 7th Int. Conf. Intell. Syst. Control (ISCON), 2025, doi: 10.1109/ISCON65210.2025.11341468. [6] H. Shi, X. Mu, and S. Wang, “BVCNN: A multi-object image recognition method based on the convolutional neural networks,” in Proc. Int. Conf. Virtual Reality Vis. (ICVRV), 2015, doi: 10.1109/ICVRV.2015.28. [7] R. Solovyev, W. Wang, and T. Gabruseva, “Weighted boxes fusion: Ensembling boxes from different object detection models,” Image Vis. Comput., vol. 107, Art. no. 104117, Mar. 2021, doi: 10.1016/j.imavis.2021.104117.
Publication Date
11/12/2026
Peer Review Type
Double-Blind Peer Review
Review Rounds: 2
Documents
Article Views
1
Total views since publication
PDF Downloads
2
Total downloads since publication
Geographic Reach
5
Digital Preservation: This article is permanently archived on Zenodo and complies with all data preservation standards across major indexing platforms.
Plagiarism Screening
This manuscript has been checked for plagiarism using TurnitinPlagiarismChecker software.
Conflict of Interest
The authors declare no conflict of interest related to this research work.
Funding Information
No funding information available for this research.
Related research in the same subject area
Loading similar articles...