Description: Field of Invention
The prime focus is to understand about abnormality detection and providing the confidence of result in Chromosome domain of various supervised methods adopted. We aim to know how the images are divided to normal and abnormal categories and increasing the confidentiality level which results in a review of an empirical study of chromosome and karyotyping, appropriate data preprocessing, experiments on different neural network systems along with conditions and comparison of them.
The Objectives of this Invention
The purpose of the invention is to enhance the accuracy of identifying the abnormal chromosome and detecting it at its earlier stages to ensure that no baby is affected from any syndrome at its earlier stages.
Background of the invention
First the study is focused on identifying chromosome then identifying its structure and then its focused on any abnormality seen on the position of chromosome leading to syndrome many scientists applied many algorithms to provide study of structure of chromosome this component will enter hospitals and biological laboratories to help the research institute perform preliminary chromosomal abnormalities investigations in early stages
In (Xingwei et al [2008]) had done with Assembling an image database involving 6900 chromosomes and implemented a genetic algorithm to optimize the topology of multi-feature based artificial neural networks (ANN). In the first layer of the scheme, a single ANN was employed to classify 24 chromosomes into seven classes. In the second layer, seven ANNs were adaptively optimized for seven classes to identify individual chromosomes. The scheme was optimized and evaluated using a “training-testing-validation” method. It results in Classification Table which include the best GA chromosome string for each of eight ANNs and the classification accuracy for both the testing and validation datasets. In the first layer of the classifier, the best GA chromosome that represents the ANN-1
First, the classification accuracy for sub-group C remains substantially lower compared to the other sub-groups of the chromosomes. Second, although we established a large and diverse database in this study, the metaphase chromosomes were all collected from normal samples. Because the numerical and structural changes can often be found in samples diagnosed with cancers and other genetic disorders, we do not know to date whether the performance of our scheme will be significantly affected when it is applied to classify abnormal chromosomes.
In (W. Zhang et al. (2018), 11th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI), pp-1-5), had investigated how applicable the proposed method is to the doctors, a metric named proportion of well classified karyotype was also designed. A result of 91.3% was achieved on this metric, indicating that the proposed classification method could be used to aid doctors in genetic disorder diagnosis. In this study, an automatic-classification method based on CNNs was proposed. The model extracts chromosome images from karyotype and output their classes. Compared with three other methods and deep learning algorithms, our method achieved a better accuracy. The experiment also shows that the method opted in paper are applicable in real life tasks. The results of this study showed that CNNs is useful in extracting features in terms of preprocessed medical images. To investigate if the proposed method is actually applicable and acceptable to doctors, The Authors proposed a new metric PWCK, which is closely related to the medical application, since doctors focus on the accuracy of individual karyotype images? The results suggest that this proposed automatic classification method can be utilized in chromosome classification, and can potentially help doctors to save a lot of time
(Ding et al [2019]) In this paper, a preprocessing model with object segmentation and feature enhancement is proposed. Combined with the framework of deep learning network, an automatic classification model for karyotype recognition of chromosomes is constructed. The preprocessing model studies the extraction of chromosome karyotype images at the pixel level and the feature enhancement of chromosome karyotype images. The model aims at providing more interpretable information for the deep learning network. In this paper, the algorithm analysis of chromosome karyotype preprocessing is carried out, the classification recognition network is built, and the detection results of the network verify the positive role of the preprocessing model.
(Qin et al [2019]) to expedite the diagnosis, we present a novel method named Varifocal-Net
for simultaneous classification of chromosomes type and polarity using deep convolution networks. The approach consists of one global-scale network (G-Net) and one local scale network (L-Net). It follows three stages. The first stage is to learn both global and local features. We extract global features and detect finer local regions via the G-Net. Results were accessed in three parts. Firstly: provide detailed evaluation results of the proposed Varifocal-Net. Second: A comparison of the proposed method with state-of-the-art methods is given. Thirdly: Presenting additional results for analyzing our performance concerning its role in clinical karyotyping workflow, the Varifocal-Net can accurately perform classification within 1 second after operators manually segment chromosomes of a cell for each patient. The karyotyping result maps it automatically generates offer the possibility for human experts to further check and correct possible misclassifications. Moreover, warnings about possible numerical abnormalities allow operators to pay extra attention to the subsequent diagnosis.
Summary of the invention
As manual genetic diagnostics is a labour-intensive and time-consuming task, developing automatic computer-assisted genetic diagnostics systems has attracted significant research interests in the last 30 years. There will be Comparisons under different conditions, and confidence score generation. The performance of the processed data will be better with end results and provide a broad mind to work in more compatible way in the Medical Field.
Detailed Description of the Invention
The identification of human chromosomes by visual identification has become a most important procedure for screening and diagnosing of various chromosomal disorders.
The most common approach to achieve the goal is Karyotyping and detecting abnormality on the Karyograms. Karyotyping is well known standardized technique used to differentiate metaphase chromosomes into 24 types called Karyograms, resulting in revealing of diverse structural changes, such as chromosomal deletions, duplications, translocations, or inversions.
As manual diagnostics requires larger man power and also takes more time for diagnosing, developing of an automatic computer self-assisting genetic diagnosis system is been keen research attracted to many researchers for the past years. To do so a Targeted model should satisfy in every aspect as shown in Figure1.2 structured which defines the flow of procedure of building a targeted system. The flow chart clearly indicating first to consider the raw data, opting for proper preprocessing method of data if the method matches the data, then at the next level, we have to select the hyperparameters to evaluate the model at different condition scenario. If not the procedure repeats in framing the dataset from the raw dataset.
After the model has reached the requirement, the confidentiality score towards accuracy has to be evaluated and there in ends with properly building up of Targeted System if not again start from scratch. The original dataset be unarranged and unstable, it is not easy to train and make model accurate, and an appropriate preprocessing method has to be considered to process the Karyogram dataset. Approaching with the probability of combining different neural network techniques to know the hyperparameters required to train the model such as learning rate, batch size and finally coming up with the best approach suiting for the hyperparameters. Several experiments need to be done with pre trained dataset which suits to all scenarios for the best clauses. Targeted Model should be more reliable with predicting the results and there should be a procedure to access the confident score of the model accuracy.

4 Claims & 1 Figure
Brief description of Drawing
In the figure which are illustrate exemplary embodiments of the invention.
Figure 1, Basic blocks to build the targeted system. , Claims: The scope of the invention is defined by the following claims:
Claim:
1. A system/method for diagnosis of wolf Hirschhorn syndrome using machine learning, said system/method comprising the steps of:
a) The data preprocessing (1) has starting point of this invention. Then, the matching of data preprocessing (2) is checked.
b) The its matched, then select the option for appropriate hyper parameter (3), if its not a appropriate hyper parameter, then search it again from the dataset (7). If its matched, evaluate the different test condition (4), the test condition score (5) is low then reconstruct the RAW data (6).
c) If the test score is high (8), we can say that the chromosome is affected.
2. As mentioned in claim 1, the data processing a start process of proposed invention. From that, the matching process is going to check, if its not matched reconstruct the RAW data.
3. As per claim 1, based the hyper parameter selection, the different test condition is evaluated with scores. If the scores are very low, then reselect the hyper parameter from the dataset.
4. As mentioned in claim 1, if test score is high, then we stop the process and concluded that, the chromosome is affected.