ORCHIDATOR: An Intelligent Orchid Recognition and Knowledge Base System Using Deep Learning

Orchids are some of the most important economic plants, costing several tens of millions of Baht in global trade each year. Thailand is the world’s leading orchid exporter, worth up to 2 billion baht per year. Currently, there are many different orchid species, differentiated by their flowers, roots, stems, and leaves. The number of natural species classified by botanists are 25,000 species, not counting the many hybrid species. It causes difficulty for the general public to recognize and identify orchid species.

 

Therefore, Dr. Wudhichart Sawangphol, Assistant Dean, together with Asst. Prof. Dr. Suppawong Tuarob, Head of Machine Intelligence and Knowledge Engineering, Miss Natnaree Sornkongdang, Miss Pimpisut Kewsuwan, and Miss Kadesuda Tantisereerat, Alumni of the Faculty of Information and Communication Technology, Mahidol University, conducted the research, ORCHIDATOR: An Intelligent Orchid Recognition and Knowledge Base System Using Deep Learning, and developed the application “ORCHIDATOR”. This research won the Recognition level of the Invention Award, under the category, Information Technology and Communication Arts from the “Invention and Innovation Contest, Fiscal year 2021”, held by the National Research Council of Thailand.

The “ORCHIDATOR: An Intelligent Orchid Recognition and Knowledge Base System Using Deep Learning” was developed to assist in the automatic identification and association of orchid data through the use of artificial intelligence (AI) technology, including the application of deep learning principles with mathematical models. This application allows users to easily and quickly search for orchid information, with an average accuracy of 95%. Users can take a photo of an orchid and upload it to the “ORCHIDATOR” application which will immediately process and identify the species of such orchid. Furthermore, the application can provide additional information of the orchid such as general information, shape, flowering season, colors, patterns, and places to buy it.

Currently, the “ORCHIDATOR” application can identify 14 popular orchid species. It is now being developed to be more precise and accurate, as well as able to identify many more Thai orchid species. Please contact Dr. Wudhichart Sawangphon at https://wudhichart.github.io/my-profile/, if you are interested in this research.