Skip to main content
ARS Home » Research » Publications at this Location » Publication #220532
Title: POULTRY SKIN TUMOR DETECTION IN HYPERSPECTRAL REFLECTANCE IMAGES BY COMBINING CLASSIFIERS

Author
item XU, CHENZHE - MYONGJI UNIV, KOREA
item KIM, INTAEK - MYONGJI UNIV, KOREA
item Kim, Moon

Submitted to: Computers in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/1/2007
Publication Date: 8/1/2007
Citation: Xu, C., Kim, I.M., Kim, M.S. 2007. Poultry skin tumor detection in hyperspectral reflectance images by combining classifiers. Computers in Agriculture. (4633):1289-12969.

Interpretive Summary: Consumptions of poultry and its products have steadily increased in recent years. In order to meet the increasing demands for safe and high quality poultry products, there is a need for the development of objective and rapid inspection methods and devices. Scientists at the Food Safety Laboratory, ARS, USDA, have been developing various nondestructive sensing methodologies and technologies, as a rapid means to inspect agricultural commodities for its safety and quality with potential online applications at the processing plants. This paper reports the use of hyperspectral imagery and development of analytical methods for detection of skin tumor spots on the chicken carcasses. Algorithms using a combination of analytical methods were devised to improve the detection of skin tumors on chicken carcasses. This information is useful to the action agency such as Food Safety Inspection Services, USDA and poultry processing plants.

Technical Abstract: This paper presents a new method for detecting poultry skin tumors in hyperspectral reflectance images. We employ the principal component analysis (PCA), discrete wavelet transform (DWT), and kernel discriminant analysis (KDA) to extract the independent feature sets in hyperspectral reflectance image data. These features are individually classified by a linear classifier and their classification results are combined using product rule. The final classification result based on the proposed method shows the better performance in detecting tumors compared with previous works.