Classification Based on LBP and SVM for Human Embryo Microscope Images

Abstract

Embryo transfer is an extremely important step in the process of invitro fertilization and embryo transfer (IVF-ET). The identification of the embryo with the greatest potential for producing a child is a very big challenge faced by embryologists. Most current scoring systems of assessing embryo viability are based on doctors' subjective visual analysis of the embryos' morphological features. So it provides only a very rough guide to potential. A classifier as a computer-aided method which is based on Pattern Recognition can help to automatically and accurately select embryos. This paper presents a classifier based on the support vector machine (SVM) algorithm. Key characteristics are formulated by using the local binary pattern (LBP) algorithm, which can eliminate the inter-observer variation, thus adding objectivity to the selection process. The experiment is done with 185 embryo images, including 47 "good" and 138 "bad" embryo images. The result shows our proposed method is robust and accurate, and the accurate rate of classification can reach about 80.42%.

Predicting pregnancy rate following multiple embryo transfers using algorithms developed through static image analysis

“…In conclusion, although the proposed algorithm may not identify which embryo or embryos will lead to a successful pregnancy, it could help to predict the pregnancy rate based on the group of embryos actually transferred, which is an important goal for IVF treatment. In our next study, we plan to select two to three embryos with high implantation potential based on our previous findings (Yin et al, 2013), and use the method proposed in the present study to make a prediction with the combination of selected embryos. In addition, a combination of this method with time lapse or pre-implantation genetic screening might provide some useful information for improving the rate of successful pregnancies and accurately predicting the probability of pregnancy.…”

“…The texture features are not based on the features of a single pixel but are the statistical calculations in an image region (containing more than one pixel). The LBP is an operator used to describe local texture features (Manna et al, 2013;Ojala et al, 1996;Yin et al, 2013), and reflects the relationship between every pixel and its surrounding pixels. The LBP is not sensitive to light, with a rotating shift-invariance.…”

“…In the present study, a novel algorithm based on pattern recognition technology is proposed for predicting the implantation potential of embryos. The algorithm combines the local binary pattern (LBP) (Manna et al, 2013;Ojala et al, 1996;Yin et al, 2013) and the Adaboost classifiers (Morales et al, 2008b), and considers all embryo images from a transfer cycle as one unit to predict pregnancy rate. The first step of the proposed method was to extract the features of the embryo images using the LBP operator.…”

Image Recommendation Algorithm Combined with Deep Neural Network Designed for Social Networks

“…After LBP and other features like RTU and BG3 are extracted, SVM is used as the classifier. e LBP is the local binary pattern as described in the paper [32], the BG3 is explained in [33], and RTU refers to remote terminal units found in [34]. e static texture recognition results are shown in Table 2 and Figure 2.…”