Abstract: The dataset of image feature for Chinese Pieridae specimen is created based on researches of butterfly image recognition. This dataset collects the color, texture and shape information of Chinese Pieridae specimen from 457 processed images by standardized feature extraction. Each record contains classification information, image identifier, feature ID and feature value of an image. The dataset can support research on pattern recognition and entomology, supplementing the data of image feature for Chinese butterfly specimen at large.
Keywords: Pieridae; specimen image; image feature; pattern recognition
|Title||A dataset of image feature for Chinese Pieridae specimen|
|Data authors||Wang Jiangning, Han Yan, Ji Liqiang|
|Data Corresponding author||Ji Liqiang (email@example.com)|
|Data volume||1.62 MB|
|Data service system||<http://www.sciencedb.cn/dataSet/handle/483>|
|Sources of funding||"Insect Image Recognition Based on Online Learning and Its Application in Butterfly Recognition" (Grant No. 31501841), National Natural Science Youth Fund (2016);"Fundamental Science Data Sharing Platform" (DKA2017-12-02-20), National R&D Infrastructure and Facility Development Program of China (2017)|
|Dataset composition||This dataset collected feature data of Chinese Papilionidae specimen from 457 images, including the color, shape, texture and other indices of the original specimen images. "Sheet1" of the Excel file consists of four data fields. For sample description of each field, readers can refer to an associated article published earlier, titled "A dataset of image feature for Chinese Papilionidae specimen".|
|Associated dataset||"A dataset of image feature for Chinese Papilionidae specimen", published in China Scientific Data (Vol.1, No.3, 2017), available at: <http://www.csdata.org/en/p/34/>.|
|Association description||This study uses the same methods for data collection, processing, quality control and assessment as the associated study, but generates data for a different butterfly taxon –Chinese Pieridae.|
Compared with other types of life, animal assumes a more complex classification system, and the collection of animal images is naturally more difficult. For butterflies, there is not enough usable data due to lack of specimen collection, image collection and hence feature extraction. To facilitate the research and application of butterfly pattern recognition, in 2016, we published the first installment of our butterfly series – "A dataset of image feature for Chinese Papilionidae specimen".1
Pieridae and Papilionidae are two common families of butterfly. As the second installment of the butterfly series, here we first collected and sorted out Pieridae samples as recorded by Monographia Rhopalocerorum Sinensium (1994 edition),2 and then used the same methods1,3 to extract the color, texture and shape information of Chinese butterflies from 457 processed images. Numerical data of the most commonly used features – color, texture and shape – were obtained. The feature data set of the Pieridae images is equipped with accurate taxonomic information and pattern recognition method.
This data set used the same methods for data collection and quality assessment as our previous study, and this set hence assumes similar data structures and usages. Specifics have been detailed in its associated study,1 which is not going to be repeated here. A brief comparison between the two studies is shown in Table 1. Altogether, these two datasets make up a more complete set of butterfly feature data. But because the acquisition of the data takes time and targets at different biological groups, they constitute relatively independent data sets, jointly supplementing the butterfly data. Entomologists can use them in combination when necessary.
Wang J, Han Y & Ji L. A dataset of image feature for Chinese Papilionidae specimen. China Scientific Data 1 (2016). DOI: 10.11922/csdata.180.2015.0008
Zhou Y. Monographia Rhopalocerorum Sinensium. Zhengzhou: Henan Science and Technology Press, 1994.
1. Wang J, Han Y & Ji L. A dataset of image feature for Chinese Pieridae specimen, Science Data Bank. DOI: 10.11922/sciencedb.483
How to cite this article
Wang J, Han Y & Ji L. A dataset of image feature for Chinese Pieridae specimen, China Scientific Data 1 (2018), DOI: 10.11922/csdata.2017.15.zh