NEURAL NETWORK WITH PRINCIPAL COMPONENT ANALYSIS FOR POULTRY CARCASS CLASSIFICATION

This paper reports the results of applying principal component analyses (PCA) of spectral reflectance data to reduce the number of input nodes for neural networks for classification of wholesome and unwholesome poultry carcasses. The results showed that the models with PCA pretreatment of input data performed better than those models without pretreatment. When sensing moving poultry carcasses in an environment without room light with a visible/near-infrared spectrophotometer, the neural network classification models with PCA pretreatment achieved 100% accuracies for training, validating, and testing. For carcasses moving at 60 birds/min, 50 factors were required for perfect classification, while for 90 birds/min 30 factors were required. When sensing in room light, the best model was generated with 30 factors for a shackle speed of 60 birds/min, with a test set accuracy of 95.8%. For 90 birds/min, the best model with a test set accuracy of 96.8% was obtained when 15 factors were used. This study showed that PCA reduced the number of input nodes to the neural network classifiers and, in most cases, improved the model's classification accuracy. It also required fewer training samples and reduced training time.     

从E制造看美国如何实现E织品质量监测

E制造,使工厂数码化网络化的方式正以因特网为依托与许多其他公司发生有趣的联系,从而导致传统行业发生深刻的变革.在美国,库特勒汉姆、蒂姆肯、福特汽车等这样的传统行业公司正在成为变革中的行业领导者.公司之间的合作已不仅是传统意义上的协作了,而是根据E制造要求和指定,设计员按统一的模式设计和生产,产供销在无形中就已完成.例如:在圣乔斯的IBM控制中心将多家海外工厂输入数据库形成编码.控制中心向芝加哥、奥地利和新加坡的仓库发出指令,客户可在数分钟内收到相关数据,从而将确定订单的型号以及样品设计和制造.又如,在明尼苏达州的公司将设计项目发给香港或其他亚洲城市的模型公司,然后将设计模型发往美国的自动装配公司,生产出北极猫履带式雪车.     

MICROBIOLOGICAL QUALITY OF BROILER CHICKEN CARCASSES PROCESSED HYGIENICALLY IN A SMALL SCALE POULTRY PROCESSING UNIT

Though large‐scale enterprises in highly‐developed countries often produce millions of items in a single day, world‐wide the volume of food produced in small‐scale enterprises or at home in countries still developing is of great economic importance and hygienic concern. Trials were conducted to determine the degree to which the sanitary state of broilers processed on a small scale could be improved through the application of simple hygienic practices. Dressed broiler chickens were examined for microbiological quality. Live birds carried high microbial load (log cfu/cm²): TPC 5.34; S. aureus 3.08 and fecal streptococci 3.02. The evisceration process resulted in a significant (P > 0.05) increase in microbial fecal contamination. Statistically significant lower microbial counts were observed in hygienically processed carcasses as compared with market carcasses. E. coli and fecal streptococci were completely absent in hygienically processed carcasses. Modest changes inprocessing methods resulted in broilers more hygienic with a gain in shelf‐life compared with those regularly available on the market.     

人工神经网络在食品工业中的应用

人工神经网络对非线性系统有很强的处理能力，适合对食品加工过程的仿真。综述了人工神经网络在加工：建模、工艺优化、过程控制与预测及在食品分析中的应用，旨在为人工神经网络在食品工业中的更广泛应用提供一定的理论基础和依据。     

EVALUATION OF ARTIFICIAL NEURAL NETWORKS FOR PINEAPPLE GRADING

The objective of this study is to evaluate resonant frequency, firmness and soluble solids for pineapple classification using artificial neural networks (ANNs) as the analytical tool. A sample of 149 pineapples was classified based on their internal qualities into five classes: unripe, partially ripe, ripe, partially overripe and completely overripe. The developed ANN model successfully classified pineapples into merely three classes as unripe, ripe and completely overripe. The most effective model was obtained when both resonant frequency and soluble solids were included in the model. The classification accuracy was more than 83% for all three classes.     

从E制造香美国如何实现E织品质量监测

E制造,使工厂数码化网络化的方式正以因特网为依托与许多其他公司发生有趣的联系,从而导致传统行业发生深刻的变革。在美国,库特勒汉姆、蒂姆肯、福特汽车等这样的传统行业公司正在成为变革中的行业领导者。公司之间的合作已不仅是传统意义上的协作了,而是根据E制造要求和指定，设计员按统一的模式设计和生产,产供销在无形中就已完成。例     

ACOUSTIC TESTING OF SNACK FOOD CRISPNESS USING NEURAL NETWORKS

The crispness of selected snack foods, Pringles potato chips, Paprika extruded snack and Munchy crackers, was evaluated by crushing the snack samples with a pair of pincers, and analyzing the frequency domain spectra of acoustic patterns by the neural networks (NNs). The inputs for training the NNs comprised 102 amplitudes of sound signals in 0–7 kHz frequency range at the intervals of about 69 Hz with moisture content or crispness grades as outputs. Both backpropagation (BPNN) and probabilistic (PNN) models showed good performance in classifying the snack foods into four grades of sensory crispness. The prediction accuracy of PNN models ranged approximately from 96 to 98% and was higher than the accuracy of BPNN models by about 10 to 25%. Results showed that frequency domain spectra of acoustic signals could be successfully analyzed by the NNs for predicting the crispness of snack food products.     

PREDICTION OF MECHANICAL PROPERTIES OF CUMIN SEED USING ARTIFICIAL NEURAL NETWORKS

In this paper, two artificial neural networks (ANNs) are applied to acquire the relationship between the mechanical properties and moisture content of cumin seed, using the data of quasi-static loading test. In establishing these relationship, the moisture content, seed size, loading rate and seed orientation were taken as the inputs of both models. The force and energy required for fracturing of cumin seed, under quasi-static loading were taken as the outputs of two models. The activation function in the output layer of models obeyed a linear output, whereas the activation function in the hidden layers were in the form of a sigmoid function. Adjusting ANN parameters such as learning rate and number of neurons and hidden layers affected the accuracy of force and energy prediction. Comparison of the predicted and experimented data showed that the ANN models used to predict the relationships of mechanical properties of cumin seed have a good learning precision and good generalization, because the root mean square errors of the predicated data by ANNs were rather low (4.6 and 7.7% for the force and energy, respectively).

PRACTICAL APPLICATIONS

Cumin seed is generally used as a food additive in the form of powder for imparting flavor to different food preparations and for a variety of medicinal properties. Physical properties of cumin seeds are essential for the design of equipment for handling, harvesting, aeration, drying, storing, grinding and processing. For powder preparation especially the fracture behavior of the seeds are essential. These properties are affected by numerous factors such as size, form and moisture content of the grain and deformation speed. A neural network model was developed that can be used to predict the relationships of mechanical properties. Artificial neural network models are powerful empirical models approach, which can be compared with mathematical models.     

Hopfield连续神经网络局部稳定性分析

利用微分几何的方法,将非线性Hopfield神经网络在平衡点附近精确线性化,使之变换成线性系统。通过分析神经网络的n阶李导数,给出并证明了Hopfield神经网络的局部稳定性定理。定理表明,当神经网络满足精确线性化条件时,系统的可控矩阵的非奇异性充分决定了神经网络吸引子的局部稳定范围。     

回归神经网络——分光光度法同时测定混合体系中多种维生素B

提出了一种将回归神经元网络与分光光度法相结合，同时测定吸收波峰重叠的维生素Ｂ类化合物（ＶＢ１，ＶＢ２，ＶＢ６，ＶＢ１２）含量的新方法，其相对误差均＜２％．     

EFFECTS OF ELECTRICAL STIMULATION ON TALLOW SUPPLEMENTED STEER CARCASSES1

Forty-nine Angus and Angus X Hereford steers were slaughtered to determine the effects of time on feed, electrical stimulation and 4% tallow in the diet on the carcass and various chemical and physical characteristics of the lean. All cattle were backgrounded on winter annuals and small grain pastures for a period of 5 months. Following the backgrounding period, ten steers were slaughtered and the rest randomly assigned to 4 treatments. The treatments consisted of high grain or grain plus 4% tallow diets, fed for either 45 or 75 days. At slaughter, the left side of each carcass was ES. Rib-fat exhibited an increase in redness with a decrease in color intensity due to electrical stimulation (P< .05). Overall, ES decreased percentage purge loss while increasing percentage cook loss (P < .05). Electrical stimulation had no significant effect on percentage thaw loss and percentage cumulative weight loss of rib steaks. Rib steaks from 75 day grain fed carcasses exhibited the greatest improvements in palatability due to ES. Across treatments the effects of ES on taste panel traits were minimal, with time on feed and diet having a greater effect on observed sensory traits than ES.     

人工神经网络用于氧化叔胺合成条件的预测及优化

对双氧水氧化法合成氧化叔胺的工艺条件，以Ｌ３（３）正交试验的结果，采用反向传播（ＢＰ）算法，完成了神经网络的学习，据此预测了反应温度、反应时间和双氧水过量比对产物得率的影响．预测结果表明，在有机酸催化作用下的最佳工艺条件为：反应温度为６８～７１℃，反应时间为１２～１４ｈ，双氧水过量１１％～１５％．从而说明了ＢＰ神经网络对合成氧化叔胺有较好的离线学习功能，可以较少的试验次数取得较多的有用信息，指导合成工艺路线