Performant barcode decoding for herbarium specimen images using vector-assisted region proposals (VARP).

Area of Science:

• Botany
• Computer Science
• Data Science

Background:

• Herbarium digitization is essential for biodiversity research but is costly and time-consuming.
• Automating post-processing tasks, such as barcode association, is crucial for efficient digitization.
• Current methods for decoding barcodes from specimen images can be slow and error-prone.

Purpose of the Study:

• To develop an automated method for accurate and fast barcode decoding from herbarium specimen images.
• To improve the efficiency of herbarium digitization workflows.

Main Methods:

• Utilizes geometric features of barcodes to identify potential barcode regions within specimen images.
• Creates a reduced composite image from identified regions for barcode decoding.
• Employs traditional barcode reading libraries for decoding.

Main Results:

• Achieved a high success rate of 96.5% in barcode decoding.
• Demonstrated a processing time of 617 ms, making it the second fastest method tested.
• Outperformed existing solutions in terms of accuracy.

Conclusions:

• The developed method significantly enhances the speed and accuracy of barcode decoding for herbarium digitization.
• This innovation supports real-time post-processing automation, crucial for large-scale digitization projects.
• The technique has potential applications beyond herbaria in other high-resolution imaging contexts.