Sara Sharifzadeh

Supervised feature selection for linear and non-linear regression of L*a*b* color from multispectral images of meat

In food quality monitoring, color is an important indicator factor of quality. The CIELab (L*a*b*) color space as a device independent color space is an appropriate means in this case. The commonly used colorimeter instruments can neither measure the L*a*b color in a wide area over the target surface nor in a contact-less mode. However, developing algorithms for conversion of food items images into L*a*b color space can solve both of these issues. This paper addresses the problem of L*a*b color prediction from multispectral images of different types of raw meat. The efficiency of using multispectral images instead of the standard RGB is investigated. In addition, it is demonstrated that due to the fiber structure and transparency of raw meat, the prediction models built on the standard color patches do not work for raw meat test samples. As a result, multispectral images of different types of meat samples (430-970nm) were used for training and testing of the L*a*b prediction models. Finding a sparse solution or the use of a minimum number of bands is of particular interest to make an industrial vision set-up simpler and cost effective. In this paper, a wide range of linear, non-linear, kernel-based regression and sparse regression methods are compared. In order to improve the prediction results of these models, we propose a supervised feature selection strategy which is compared with the Principal component analysis (PCA) as a pre-processing step. The results showed that the proposed feature selection method outperforms the PCA for both linear and non-linear methods. The highest performance was obtained by linear ridge regression applied on the selected features from the proposed Elastic net (EN) -based feature selection strategy. All the best models use a reduced number of wavelengths for each of the L*a*b components.

Sparse supervised principal component analysis (SSPCA) for dimension reduction and variable selection

Abstract Principal component analysis (PCA) 1 is one of the main unsupervised pre-processing methods for dimension reduction. When the training labels are available, it is worth using a supervised PCA strategy. In cases that both dimension reduction and variable selection are required, sparse PCA (SPCA) methods are preferred. In this paper, a sparse supervised PCA (SSPCA) method is proposed for pre-processing. This method is appropriate especially in problems where, a high dimensional input necessitates the use of a sparse method and a target label is also available to guide the variable selection strategy. Such a method is valuable in many Engineering and scientific problems, when the number of training samples is also limited. The Hilbert Schmidt Independence Criteria (HSIC) is used to form an objective based on minimization of a loss function and an L 1 norm is used for regularization of the Eigen vectors. While the proposed objective function allows a sparse low rank solution for both linear and non-linear relationships between the input and response matrices, other similar methods in this case are only based on a linear model. The objective is solved based on penalized matrix decomposition (PMD) algorithm. We compare the proposed method with PCA, PMD-based SPCA and supervised PCA. In addition, SSPCA is also compared with sparse partial least squares (SPLS), due to the similarity between the two objective functions. Experimental results from the simulated as well as real data sets show that, SSPCA provides an appropriate trade-off between accuracy and sparsity. Comparisons show that, in terms of sparsity, SSPCA performs the highest level of variable reduction and also, in terms of accuracy it is one of the most successful methods. Therefore, the Eigen vectors found by SSPCA can be used for feature selection in various high dimensional problems.

Statistical Quality Assessment of Pre-fried Carrots Using Multispectral Imaging

Multispectral imaging is increasingly being used for quality assessment of food items due to its non-invasive benefits. In this paper, we investigate the use of multispectral images of pre-fried carrots, to detect changes over a period of 14 days. The idea is to distinguish changes in quality from spectral images of visible and NIR bands. High dimensional feature vectors were formed from all possible ratios of spectral bands in 9 different percentiles per piece of carrot. We propose to use a multiple hypothesis testing technique based on the Benjamini-Hachberg (BH) method to distinguish possible significant changes in features during the inspection days. Discrimination by the SVM classifier supported these results. Additionally, 2-sided t-tests on the predictions of the elastic-net regressions were carried out to compare our results with previous studies on fried carrots. The experimental results showed that the most significant changes occured in day 2 and day 14.

Optimal vision system design for characterization of apples using US/VIS/NIR spectroscopy data

Quality monitoring of the food items by spectroscopy provides information in a large number of wavelengths including highly correlated and redundant information. Although increasing the information, the increase in the number of wavelengths causes the vision set-up to be more complex and expensive. In this paper, three sparse regression methods; lasso, elastic-net and fused lasso are employed for estimation of the chemical and physical characteristics of one apple cultivar using their high dimensional spectroscopic measurements. The use of sparse regression reduces the number of required wavelengths for prediction and thus, simplifies the required vision set-up. It is shown that, considering a tradeoff between the number of selected bands and the corresponding validation performance during the training step can result in a significant reduction in the number of bands at a small price in the test performance. Furthermore, appropriate regression methods for different number of bands and spectrophotometer design are determined.

Learning industrial robot force/torque compensation: A comparison of support vector and random forests regression

Haptics, as well as force and torque measurements, are increasingly gaining attention in the fields of kinesthetic learning and robot Learning from demonstration (LfD). For such learning techniques, it is essential to obtain accurate force and torque measurements in order to enable accurate control. However, force and torque measurements using a 6-axis force and torque sensor mounted at the end effector of an industrial robot are known to be corrupted due to the robots internal forces, gravity, un-modelled dynamics and nonlinear effects. This paper presents an evaluation of two techniques, SVR and Random Forests, to recover the external forces and accurately selected possible contact situations by estimating a robots internal forces. The performance of the learned models have been evaluated using different performance metrics and comparing them with respect to the features contained in the input space. Both SVR and Random Forests require low computational complexity without intensive training over the operational space under the given ssumptions. In addition, these methods do not need data to be available online. The SVR and Random Forests models are experimentally validated using Motoman SDA10D dual-arm industrial robot controlled by Robot Operating System (ROS). The experiments showed that force and torque compensation based on Random Forests has outperformed Support Vector Regression.

DCT-based characterization of milk products using diffuse reflectance images

We propose to use the two-dimensional Discrete Cosine Transform (DCT) for decomposition of diffuse reflectance images of laser illumination on milk products in different wavelengths. Based on the prior knowledge about the characteristics of the images, the initial feature vectors are formed at each wavelength. The low order DCT coefficients are used to quantify the optical properties. In addition, the entropy information of the higher order DCT coefficients is used to include the illumination interference effects near the incident point. The discrimination powers of the features are computed and used to do wavelength and feature selection. Using the selected features of just one band, we could characterize and discriminate eight different milk products. Comparing this result with the current characterization method based of a fitted log-log linear model, shows that the proposed method can discriminate milk from yogurt products better.