Many publicly available barcode image sets have appeared online. In most cases, publicly available barcode image sets are used to train neural networks or to test the barcode recognition performance of a barcode reader.
VintaSoft is the author of the
VintaSoft Barcode .NET SDK, a professional cross-platform SDK for Windows, Linux, and macOS that enables recognition and generation of 1D and 2D barcodes in .NET, WPF and Web. We test the
VintaSoft Barcode .NET SDK using our test system, which contains approximately 17,000 images and over 20,000 barcodes. Our test system is not publicly available, so we can publish information about our test system, but we cannot provide access to it. Some skeptics might say that all this is just talk without proof.
In this article, we will test
VintaSoft Barcode .NET SDK on publicly available sets of barcode images, publish the test results, and provide .NET projects that anyone can use to replicate the test results.
As public image sets with barcodes we will take image sets
Artelab Medium Barcode 1D Collection and
DEAL KAIST Lab Barcode Main.
Recognizing barcodes in the set of images "Artelab Medium Barcode 1D Collection" using VintaSoft Barcode .NET SDK
The "Artelab Medium Barcode 1D Collection" image set contains 430 images with EAN13 barcodes and can be downloaded from the following link:
http://artelab.dista.uninsubria.it/downloads/datasets/barcode/medium_barcode_1d/medium_barcode_1d.html
We have created a console .NET project for recognizing barcodes in the "Artelab Medium Barcode 1D Collection" image set using
VintaSoft Barcode .NET SDK 16.0.1.1.
Here is the result of a .NET console project work for recognizing barcodes in a set of images "Artelab Medium Barcode 1D Collection":
========================================
Result: VintaSoft Barcode .NET SDK
========================================
Total number of files: 430
Total recognition time: 37942 ms
Correctly recognized: 382 (89%)
Recognition time per image: 88 ms
========================================
ZIP-archive with .NET-project can be
downloaded here.
If you want to reproduce our test by yourself, you need to complete the following steps:
- Download the ZIP archive with our .NET project.
- Download the "Artelab Medium Barcode 1D Collection" image set.
- Run the project on the folder containing the image set and see the result.
Recognizing barcodes in the set of images "DEAL KAIST Lab Barcode Main" using VintaSoft Barcode .NET SDK
The "DEAL KAIST Lab Barcode Main" image set contains 3200 images with EAN13 barcodes and can be downloaded from the following link:
https://www.kaggle.com/datasets/s0dium/deal-kaist-lab-barcode-main
We have created a console .NET project for recognizing barcodes in the "DEAL KAIST Lab Barcode Main" image set using
VintaSoft Barcode .NET SDK 16.0.1.1.
Here is the result of a .NET console project work for recognizing barcodes in a set of images "DEAL KAIST Lab Barcode Main":
========================================
Result: VintaSoft Barcode .NET SDK
========================================
Total number of files: 3200
Total recognition time: 245968 ms
Correctly recognized: 2827 (88%)
Recognition time per image: 77 ms
========================================
ZIP-archive with .NET-project can be
downloaded here.
If you want to reproduce our test by yourself, you need to complete the following steps:
- Download the ZIP archive with our .NET project.
- Download the "DEAL KAIST Lab Barcode Main" image set.
- Run the project on the folder containing the image set and see the result.
Test results for VintaSoft Barcode .NET SDK on publicly available barcode image sets
We analyzed the tested images and realized that
VintaSoft Barcode .NET SDK was unable to recognize some barcodes because barcode image has low quality. Here is an example of barcode image with low quality:
We consider the test results near to outstanding. At the same time, we are constantly improving our algorithms and are not resting on our laurels.
If you would like us to test
VintaSoft Barcode .NET SDK on another publicly available image set with barcodes, please
contact us.
