Fruit detection with binary partition trees (İkili bölüntü ağaçları ile meyve tespiti)

Official URL: http://dx.doi.org/10.1109/SIU.2016.7496082

In this study, binary partition trees are applied to the problem of fruit detection. The fact that binary partition trees are inherently unbiased and independent of flatzones is the main reason for this application. Using only circularity from the shape priors, this system is put to test with 39 images of three classes of fruits and the test results show an average of 0.669 precision and 0.851 recall.