A neuro-fuzzy based image classification system that utilizes color-imaging
features of poultry viscera in the spectral and spatial domains was develo
ped in this study. Color images of 320 livers and hearts from normal, airsa
cculitis, cadaver and septicemia chickens were collected in the poultry pro
cess plant These images in red, green, and blue (RGB) color space were segm
ented and statistical analysis was performed for feature selection. A neuro
-fuzzy system utilizing hybrid paradigms of fuzzy inference system and neur
al networks was used to enhance the robustness of the classification proces
ses. The accuracy for separation of normal from abnormal livers ranged 87.5
to 92.5%, when two classes of validation data were used. For classificatio
n of normal and abnormal chicken hearts, the accuracies were 92.5 to 97.5%.
When neuro-fuzzy models were employed to separate chicken livers into norm
al, airsacculitis, and cadaver the accuracy was 88.3% for the training data
and 83.3% for the validation data. Combining features of chicken liver and
heart, a generalized neuro-fuzzy model was designed to classify poultry vi
scera into four classes (normal, airsacculitis, cadaver and septicemia). Th
e classification accuracy was 86.3% for training and 82.5% for validation.