Blooming flower recognition by using eigenvalues of shape features

Wooi-Nee Tan; Racheal Sem; Yi-Fei Tan

doi:10.1117/12.2064504

Translator Disclaimer

16 April 2014 Blooming flower recognition by using eigenvalues of shape features

Proceedings Volume 9159, Sixth International Conference on Digital Image Processing (ICDIP 2014); 91591R (2014) https://doi.org/10.1117/12.2064504
Event: Sixth International Conference on Digital Image Processing, 2014, Athens, Greece

Abstract

This paper introduces the concept of eigenvalues in describing the shape features of blooming flowers, and implements the idea in recognizing the blooming flowers automatically. The input images of blooming flowers are taken from natural scene in the form of RGB images. The proposed method first segments and crops the targeted flower object, then calculates four shape features to form a 2 × 2 matrix. Eigenvalues of the matrix computed from the testing set are then used to compare with those eigenvalues of the reference set. The advantage of utilizing the idea of eigenvalues is that the dimension of parameters used in comparison can be reduced. Based on the experimental result on database which consists of 5 types of flowers with a total of 46 images, a recognition rate of 80.43% is achieved.

Citation Download Citation

Wooi-Nee Tan, Racheal Sem, and Yi-Fei Tan "Blooming flower recognition by using eigenvalues of shape features", Proc. SPIE 9159, Sixth International Conference on Digital Image Processing (ICDIP 2014), 91591R (16 April 2014); https://doi.org/10.1117/12.2064504

ACCESS THE FULL ARTICLE