In this paper we present a new approach to categorize the wear of cutting tools used in edge prof... more In this paper we present a new approach to categorize the wear of cutting tools used in edge profile milling processes. It is based on machine learning and computer vision techniques, specifically using B-ORCHIZ, a novel shape based descriptor computed from the wear region image. A new Insert dataset with 212 images of tool wear has been created to evaluate our approach. It contains two subsets: one with images of the main cutting edge, and the other one with the edges that converge to it (called Insert-C and Insert-I, respectively). The experiments were conducted trying to discriminate between two (Low-High) and three (Low-Medium-High) different wear levels and the classification stage was carried out using a Support Vector Machine (SVM). Results show that B-ORCHIZ outperforms other shape descriptors (aZIBO and ZMEG) achieving accuracy values between 80.24% and 88.46% in the different scenarios evaluated. Moreover, a hierarchical cluster analysis was performed, offering prototype images for wear levels, which may help researchers and technicians to understand how the wear process evolves. These results show a very promising opportunity for wear monitoring automation in edge profile milling processes.
ABSTRACT Classifying damaged-intact cells in a semen sample presents the peculiarity that the tes... more ABSTRACT Classifying damaged-intact cells in a semen sample presents the peculiarity that the test class distribution is unknown. This paper studies under which design conditions the misclassification rate is minimum for the uncertainty region of interest (ratio of damaged cells lower than 20%) and (b) deals with quantifying the proportion of damaged/intact sperm cells in a given sample based on computer vision and supervised learning. We have applied a discrete wavelet transform to the spermatozoa head images and computed the mean and standard deviation (WSF) and four Haralick descriptors (WCF). Using a backpropagation neural network, the error rate averaged over distributions in the region of interest is 4.85% with WCF. The assessment of several quantification methods shows the conditions under which the Adjusted Count method leads to an overall mean absolute error of 3.2 and the Classify & Count method yields 2.4, both with WCF features. Deviations of this order are considered reasonable for this field.
Advances in image analysis make possible the automatic semen analysis in the veterinary practice.... more Advances in image analysis make possible the automatic semen analysis in the veterinary practice. The proportion of sperm cells with damaged/intact acrosome, a major aspect in this assessment, depends strongly on several factors, including animal diversity and ...
In this work a backpropagation neural network has been used in order to determine the acrosome st... more In this work a backpropagation neural network has been used in order to determine the acrosome state of boar spermatozoa. The way in which the previous image process using different mother wavelets, affects to the classi-fication success rate has been evaluated. We assessed five different wavelet families – or mother functions – : Dauchebies, Coiflets, Symlets, Meyer and biorthogonal applied to images, and then first and second order texture statistical descriptors have been obtained from the obtained coefficients. The classification has been carried out with a neural network, in which the perfor-mance obtained using different configurations have been evaluated. Results show that the hit rates of different families used vary as far as 7%. Best re-sults are obtained using biorthogonal and Symlets families, with about 95% success rate.
In this paper we present a new approach to categorize the wear of cutting tools used in edge prof... more In this paper we present a new approach to categorize the wear of cutting tools used in edge profile milling processes. It is based on machine learning and computer vision techniques, specifically using B-ORCHIZ, a novel shape based descriptor computed from the wear region image. A new Insert dataset with 212 images of tool wear has been created to evaluate our approach. It contains two subsets: one with images of the main cutting edge, and the other one with the edges that converge to it (called Insert-C and Insert-I, respectively). The experiments were conducted trying to discriminate between two (Low-High) and three (Low-Medium-High) different wear levels and the classification stage was carried out using a Support Vector Machine (SVM). Results show that B-ORCHIZ outperforms other shape descriptors (aZIBO and ZMEG) achieving accuracy values between 80.24% and 88.46% in the different scenarios evaluated. Moreover, a hierarchical cluster analysis was performed, offering prototype images for wear levels, which may help researchers and technicians to understand how the wear process evolves. These results show a very promising opportunity for wear monitoring automation in edge profile milling processes.
ABSTRACT Classifying damaged-intact cells in a semen sample presents the peculiarity that the tes... more ABSTRACT Classifying damaged-intact cells in a semen sample presents the peculiarity that the test class distribution is unknown. This paper studies under which design conditions the misclassification rate is minimum for the uncertainty region of interest (ratio of damaged cells lower than 20%) and (b) deals with quantifying the proportion of damaged/intact sperm cells in a given sample based on computer vision and supervised learning. We have applied a discrete wavelet transform to the spermatozoa head images and computed the mean and standard deviation (WSF) and four Haralick descriptors (WCF). Using a backpropagation neural network, the error rate averaged over distributions in the region of interest is 4.85% with WCF. The assessment of several quantification methods shows the conditions under which the Adjusted Count method leads to an overall mean absolute error of 3.2 and the Classify & Count method yields 2.4, both with WCF features. Deviations of this order are considered reasonable for this field.
Advances in image analysis make possible the automatic semen analysis in the veterinary practice.... more Advances in image analysis make possible the automatic semen analysis in the veterinary practice. The proportion of sperm cells with damaged/intact acrosome, a major aspect in this assessment, depends strongly on several factors, including animal diversity and ...
In this work a backpropagation neural network has been used in order to determine the acrosome st... more In this work a backpropagation neural network has been used in order to determine the acrosome state of boar spermatozoa. The way in which the previous image process using different mother wavelets, affects to the classi-fication success rate has been evaluated. We assessed five different wavelet families – or mother functions – : Dauchebies, Coiflets, Symlets, Meyer and biorthogonal applied to images, and then first and second order texture statistical descriptors have been obtained from the obtained coefficients. The classification has been carried out with a neural network, in which the perfor-mance obtained using different configurations have been evaluated. Results show that the hit rates of different families used vary as far as 7%. Best re-sults are obtained using biorthogonal and Symlets families, with about 95% success rate.
Uploads
Papers by Enrique Alegre