Chicken embryo development detection using Self-Organizing Maps and K-mean Clustering

This paper presents a set of procedures for detecting the primary embryo development of chicken eggs using Self-Organizing Mapping (SOM) technique and K-means clustering algorithm. Our strategy consists of preprocessing of an acquired color image with color space transformation, grouping the data by Self-Organizing Mapping technique and predicting the embryo development by K-means clustering method. In our experiment, the results show that our method is more efficient. Processing with this algorithm can indicate the period of chicken embryo in on hatching. By the accuracy of the algorithm depends on the adjustment the optimum number of iterative learning. For experiment the learning rate using the example of number 4 eggs, found that the optimum learning rate to be in the range of 0.1 to 0.5. And efficiency the optimum number of iterative learning to be in the range of 250 to 300 rounds.