Lung Cancer Detection Using Neural Net

Abstract

Leukemia is a sort of blood threatening development which happens due touncommon development in WBCs (white platelets) in bone marrow of human body. Leukemia can be appointed serious leukemia and progressing leukemia, affected severely and causes malignant growth. In the vast majority of the cases early in which serious leukemia turns out to be astoundingly fast however steady leukemia creates slowly .Among all kinds of kinds of developments, cell breakdown in the lungs is the most overwhelming contamination having the most critical passing rate. Dealt with tomography investigates are utilized for ID of lung hurt as it gives unmistakable image of harmful development in the body and tracks its development. Picture dealing with systems are utilized generally in clinical fields for beginning stage affirmation of cell breakdown in the lungs. The calculation for cell breakdown in the lungs disclosure is proposed utilizing strategies, for example, focus disconnecting for picture pre-dealing with followed by division of lung area of interest utilizing numerical morphological endeavors. Mathematical parts are dealt with from the eliminated district of interest and used to bundle CT channel pictures into typical and abnormal by utilizing support vector machine. Further both the sorts have two sub arrangements lymphocytic and myeloid. In this paper, we will explore different picture dealing with and Man-made intelligence procedures used for request of leukemia area and endeavor to focus in on advantages and limitations of different similar investigates to summarize an result which will be valuable for various experts.

Introduction

I. INTRODUCTION

The most popular applications of image processing are for the early diagnosis of various medical conditions. Given that it falls under the category of blood cancers, leukaemia is among the most fascinating topics of research. kind of blood cancer that can afflict individuals of all ages from infants to the elderly. Classification is quite simple because to the usage of image processing with computer-based algorithms. When leukaemia disease identification is carried out by a professional with specialised knowledge, there may be errors made as a  result of ignorance or inaccurate information found in  the microscopic image. In order to improve the detection accuracy within that field, browser algorithms can be very helpful [1]. Human blood has two different forms of WBCs, and when lymphoblastic (ALL) classification [2]. Four major  categories—ALL, AML, CLL, and CML—can be used to classify blood cancers. All of these types depend on how they affect the white blood cells (WBCs) in human blood. Initial

1. Acute Myeloid Leukemia (AML): This specific type of malignancy is brought on by poor bone marrow growth or other factors. When the WBCs increase quickly, the bone marrow's function is negatively impacted, which leads to cancer. In the majority of cases, early discovery of this type of cancer can result in a successful outcome. One may experience breathing difficulties, haemorrhage , and other symptoms during this type of malignancy [1, 3].
2. Acute Lymphocytic Leukemia (ALL): This type of malignancy is typically encountered in children, with the rapid proliferation of white blood cells being a major contributing factor. Radiation exposure, viral infections, and transmitted disorders like Down's condition are a couple of the causes connected to the ongoing sort of malignant growth. At the point when this sort of leukemia is found in more seasoned grown-ups, youngsters have a higher reduction rate than do they. Everything is partitioned into L1, L2, and L3 classes.
3. Chronic Myeloid Leukemia (CML): This type of malignancy develops when the myeloid cells experience a genetic alteration. Traditional cells cannot effectively fight infections when there is genetic change in the cells. As a slow-growing form of cancer, this type of leukaemia is prevalent in adults. The chronic phase, accelerated phase, and blast process are the additional three stages of the CML malignancy. Cancer is curable at this point since it is still in its early stages and is growing extremely slowly. It grows increasingly effective in the second phase, begins to destroy the blood cells, and then explodes in the last stage [2, 3].
4. Chronic Lymphocytic Leukemia (CLL): This variety of leukaemia affects both the myeloid and blood cells. The white blood cell count will rise with this type of malignancy, but they will not function correctly. This may have the finest likelihood of survival if somebody were to contract the disease. When it comes to adults, it is more common than when it comes to kids. Leukemic cells with B-lymphocyte properties that have grown clonally and accumulated in CLL are its distinguishing features. People who receive a CLL diagnosis occasionally develop ALL. [1, 4]

Non-little cell disintegration in the body has been the most prevalent. lungs, which account for 80–85% of all instances, whereas only 15%–20% of cases of detrimental development are treated by minimal cell breakdown in the lungs. Overall stage of lung risk is determined either by lung infection's lung spread and the size of the tumour . According to the actual world, the disintegration of lung cells typically occurs in four stages: Stage I is a lung-specific illness, Stage II and III are chest-specific dangers, and Stage IV is a lung-specific cell breakdown that has moved from the chest to various parts of the body. Different imaging modalities, such like Positron Emission Tomography (PET), Appealing Resonance Imaging (X-beam), Figured Tomography (CT), and Chest X-rays, should be able to investigate cell breakdown in the lungs. X-ray check images were taken by They are more dependable, have higher clarity, and cause less mutilation than other modalities, they are generally preferred. Visual understanding of informational collecting is a somewhat drawn-out, time-consuming, and highly individualised process. This raises the likelihood of human error and could lead to an infection being misclassified. Following that, a robotized system is of utmost essential to coordinate the radiologist in adequate assurance of cell breakdown in the lungs. This system's technique combines dataset combination, pre-handling, lung segmentation, feature extraction, and depiction. The disease starts in the lung and is caused by cell disintegration. It is the second most frequent illness among people. When symptoms are clearly visible, testing for lung cell breakdown are performed to determine the type of lung cell breakdown and whether it has spread. This process of risk metastasizing is referred to as metastasis. about 42,000 individuals 2010 is still up in the air to have cell breakdown in the lungs, which is roughly 115 persons. Over half a century ago, that connection between tobacco and risk was established. Similar to how smoking is by far the biggest risk factor for lung cell disintegration. That risk rises as daily cigarette consumption increases.

Receptivity to artificial substances or other elements in the environment, such as pollution, may increase also a chance of danger There are many billions of tiny cells in the body. Generally speaking, as cells suffer or deteriorate, they die and are the substitution of fresh cells. Cells frequently continue to divide and produce when they aren't needed, leading to a strange development known as a malignancy. Non small cell breakdown and tiny cell breakdown are the two primary types of cell breakdown in the lungs. As for the stages, there are generally four times when breathing could be dangerous for the lungs: I through. At this time, the cell breakdown in the lungs should be easier to detect and diagnose with processed tomography (CT) than with a straightforward chest X-shaft. There are two types of malignancy: dangerous disease and benign development. By splitting from the outstanding malignancy, dangerous cells propagate. Figure 1 depicts the structure of a common cell and the process by which cancer is created. Treatments for dangerous development aim to eradicate or suppress cancerous cells. Despite the fact that chemotherapy, radiation therapy, and surgery have all been utilised to treat lung cell breakdown, the combined long-term survival percentage for all phases is only 14%. The general pre-getting ready and improvement techniques have been described by B.V. Ginneken. He has divided the lung area extraction techniques into two distinct classes: rule-based and pixel-based classes.

Thresholding, region creation, edge acknowledgment, and morphologic action are employed techniques [1]. For division, the mathematical morphology's watershed computation is astonishing. has described a few of the division tactics that are needed in this context [8]. D gives still another technique for modified acknowledgment of lung handling. Lin [7]. Wiener channel also provides higher benefits for technological upgrading. Unambiguously marking split sections is a Sobel edge revealing methodology. Area of interest, calcification, grip size and condition, edge, and eccentricity are among the accepted configurations to be isolated [3, 4]. The goal with this effort is to use an image preparation process to much more correctly pinpoint the time of lung infection onset. The NIH NCI Lung image Data set Consortium (LIDC) dataset is used for the proposed work's experimentation since it provides the opportunity to do the inquiry. With death rates steadily rising, harmful development is a major global public health concern. cell death in the lungs is one example The most unmistakable and serious damage types are those that actually influence individuals. The course of perilous chest improvements (hazardous handles) brought about by unrestrained cell expansion of lung tissues is known as cell disintegration in the lungs, otherwise called carcinoma. The two principal risk factors for creating carcinogenic lung cancers are smoking and tobacco use.

With a life expectancy of around 5-6 years, the constancy pace of lung unsafe improvement victims joining all stages is an exceptionally unprecedented 14%. The fundamental issue with cell breaking down in the lungs is that most of these injury cases are found after later phases of disease, making medicine more unsafe and basically bringing down the possibilities of endurance. Accordingly, the lungs' area of cell breakdown in its previous stages can By providing the patients with sufficient rewards, you can increase their odds of sticking with it by 60–70%. swift treatment, and in doing so, it monitors the mortality rate. Little cell lung disease and non-little colony lung membrane destruction are the two main types of lung cell destruction processes that depend on cell characteristics. The non-little unit degradation in the lungs, which accounts for around 80–85% of any and all occasions, happens most often happening; by and by, a little cell breaking down inside this lungs tends to around 15% of instances of hazardous turn of events.

The association of lung-harming improvement relies upon the pace of contamination transmission and the degree of the turn of events. To guarantee truthfulness, cell breakdown in the lungs is basically mentioned in 4 phases: Stage I diseases are restricted towards the lung, Stage II and III malignancies are constrained to the chest, and Stage IV lung-threating diseases have disseminated from either the chest to numerous parts of the body. Using various imaging modalities, such as Positron Emanation Tomography (PET), Attractive Reverberation Imaging (X-ray), Figured Tomography (CT), and Chest X-radiates, cell breakdown in the lungs inspection should be possible. Since they are more dependable, have superior clarity, and are less contorted than other modalities, CT inspection images are generally preferred. An unfavourable method of information gathering includes visual interpretation, which takes a lot of time and depends heavily on the individual using it. This provides a high risk of human error and may cause disease classification error. From this point on, a motorised system is of utmost necessity for coordinating with radiology in accurate lung cell breakdown diagnosis. The method designed for this structure integrates morphological operations, dataset variety, lung division, pre-treatment, and planning.

II. LITERATURE REVIEW

Fostering a mechanized cell break in the lungs locale framework has been the subject of different examinations. For division, S. A. Patil et al. utilized morphological exercises and district making. By request to request genuine film deterioration inside this lungs, eliminated mathematical plans and starting sales authentic surface segments were applied to ANN [5]. A mechanized sharp design was set up by Amjed et al. [6] to deal with affirmation and solicitation of cell breakdown in the lungs in CT pictures. They involved mathematical components and morphological picture arranging systems in their work. A division assessment receptive to granular dealing with guess was proposed by Gathering Xie et al. For the benefit from the underlying venture, manual venturing was utilized in this assessment, and the division was finished utilizing four mathematical surface parts [7]. The division problem was redesigned in [8] as an energy minimization problem using a methodology that took Boylov's chart cut strategy into account. The identification of the lung handles seen in PET/CT images was done using watershed change and image arrangement [9]. A variety of specialists used different representation computations and first- and second-solicitation measurable surface boundaries to remove features from X-Bar images [10–12]. The arrangement of the lung handles in [11] was done via back multiplication and direct vector quantization (LVQ) relationships. Direct Discriminant Assessment (LDA) and Non-Direct Discriminant Assessment (NDA) approaches were applied for feature diminishes Head Part Assessment (PCA) in [12] Fewer than a handful of methods [13, 14] with different presentation techniques used Haralick surface portions to extract features from lung CT images. A unique scale-based image was employed by Balaji et al. Ada et al. involved controlled feed forward back spread brain association for their game system [14]. To portray the advancement on CT pictures, Mir Rayat et al. utilized the change pursue calculation to recognize perceivable parts and the Chi square distance measure [15]. Support Vector Machine (SVM) was utilized to set up a supposition model in exactly on time area of cell breakdown in the lungs after surface parts of CT pictures were taken out utilizing curvelet alteration [16]. Moderate forward choice gauge was used by Sandeep et al. in their review to choose the discriminative game plans among the GLCM surface parts.

For the gathering, multinomial multivariate Bayesian was favored [17]. Vinod Kumar and his colleagues [18] proposed a novel method for cell disclosure in the lungs using the fake neural association (ANN), cushy min-max neural association, and FCM. In light of this, we provide a strategy for individualised lung cell breakdown disclosure in PET/CT images using real surface layouts and FCM classifier. With death rates consistently increasing, improvement is representing a serious danger to worldwide wellbeing. The most obvious and hazardous kind of harm that happens in the two people and creatures is cell crumbling in the lungs. In view of unrestrained cell development in lung tissues, cell crumbling in the lungs and furthermore acknowledged carcinoma is plan of risky lung changes (threatening handles). The two essential gamble factors for creating harmful lung handles are smoking and consuming tobacco. With a range of season of around 5-6 years, the perseverance pace of cellular breakdown in the lungs patients entering all stages is very low, roughly 14%.

The essential issue with lung cell breakdown is that most of these malignant growth patients are analyzed in later stages. of affliction making drugs less secure and at last bringing down the possibilities of perseverance. Along these lines, treating patients rapidly and successfully can expand their possibilities enduring cell breakdown in the lungs in its beginning phases by 60-70%, which brings down the death rate.

The two primary sorts of cell breakdown in the lungs that rely upon cell qualities are little cell lung sickness and non-little cell breakdown in the lungs. The non-little cell breakdown in the lungs, which represents 80-85% of all occurrences, is the most regular; notwithstanding, little cell breakdown in the lungs just influences 15-20% of instances of dangerous turn of events. lung carcinogenic growth organizing depends on developing size and lung ailment spread.

To be earnest, cell breakdown in the lungs is fundamentally separated into 4 stages: Stage I cellular breakdown in the lungs is restricted to the lung, Stage II and III chest disease is obliged to the chest, and Stage IV cellular breakdown in the lungs has spread from the chest to different pieces of the body. Different imaging methods, like Positron Emanation Tomography (PET), Attractive Reverberation Imaging (X-ray), Processed Tomography (CT), and Chest X-radiates, ought to have the option to examine cell breakdown in the lungs. Since CT assessment pictures are more solid, more clear, and less wound than those got utilizing different modalities, they are frequently liked. A monotonous, tedious, and vigorously subordinate technique is visual information interpretation. incorrect disease classification Consequently, an automated framework is of utmost importance to guide the radiologist correctly identifies cell disintegration in the lungs. This framework's strategy includes data collection, pre-handling, lung division, highlight extraction, and arrangement. Some analysts have suggested and actually carried out all the recognition of genetic breakdown in the lungs using various methods for image processing and AI. A model that provides structure to the arrangement of knobs and common lung life processes was put out by Aggarwal, Furquan, and Kalra [4]. The tactic distinguishes between numerical, factual, and low level qualities. LDA is used to determine the best thresholding for division and as a classifier. The framework features 53.33% particularity, 97.14% affectability, and 84% precision. Although the framework can identify the disease knob, its accuracy is still inadequate. No AI systems have ever been used for simple division or ordering operations. Therefore, combining any of its methods into our new model doesn't increase the likelihood of advancement. Convolution neural organisation was used by Jin, Zhang, and Jin [5] as a classifier. in his PC helped plan system to perceive the cell breakdown in the lungs. The structure has 84.6% of precision, 82.5% of affectability and 86.7% of distinction.

The potential gain of this model is that it uses indirect channel in Area of premium (return for capital contributed) extraction Some analysts have suggested and actually carried out all the recognition of genetic breakdown in the lungs using various methods for image processing and AI. A model that provides structure to the arrangement of knobs and common lung life processes was put out by Aggarwal, Furquan, and Kalra [4]. The tactic distinguishes between numerical, factual, and low level qualities. LDA is used to determine the best thresholding for division and as a classifier. The framework features 53.33% particularity, 97.14% affectability, and 84% precision. Although the framework can identify the disease knob, its accuracy is still inadequate. No AI systems have ever been used for simple division or ordering operations. Therefore, combining any of its methods into our new model doesn't increase the likelihood of advancement. Convolution neural organisation was used by Jin, Zhang, and Jin [5] as a classifier. To help picture contrast, this construction utilizes unobtrusive change. Preceding division, an image is binarized, and the sections are made utilizing a unique shape model. Delicate determination approach is utilized to perform disease demands. To construct the classifier, factors including region, mean, entropy, relationship, tremendous turn length, and minor focus point length are disconnected. The design's general precision is 94.12%. Despite its blockage, the condition isn't depicted as being harmless or as not having the capacity to subvert any piece of the recommended plan. An idea that utilizes watershed division was embraced by Ignatious and Joseph [8]. It utilizes the Gabor channel during pre-cleaning to further develop the image quality. It isolates the area making process from the model's precision and mental solace. Precision of the proposed is with Interest (ROI). Area, whimsicalness, circularity, and fractal estimate are shape properties; mean, change

To prepare and define the help vector with machining and determine if the handle is benign or undermines, energy, entropy, skewness, separation, and flawlessness are isolated. This paradigm has the advantage of characterising illness as harmless or debilitating, but it has a drawback in that it requires prior knowledge of the area of interest. Our new model may benefit from the support vector machine's innocuous or undermining game plan.

According to research on writing reviews, the Ignatious and Joseph [8] structure is the best one in terms of accuracy and the benefits of the methods utilised. Pictured setting it up beforehand employs the Gabor channel to update the image, a marker-controlled watershed process for division, and a sickness manage perception. Additionally, this model eliminates the district, edge, and arbitrary characteristics that are present only in the infection handles. It compares the evaluation to other recently suggested models and emphasises that its precision, at 90.1%, is higher than those models. Even while the system is now the optimal course of action (fig. 1), it does have a few drawbacks. They are mentioned below. A small number of arrangements have been set aside for illness management No prior processing, such as turmoil removal or picture smoothing, which may help to increase the handle area, has been carried out. There has never been a gathering as harmless or as destructive of an abandoned risky development. suggested model Changes have been made to the best plan currently in place, and a new model has been suggested as seen in figure 2. Center channel and Gaussian channel have been finished in the pre-getting ready stage rather than Gabor Channel. The pre-arranged picture is divided using watershed division after pre-treatment. This produces a picture with checked illness handles. For the identified threatening development handles, features like Centroid, Distance across, and pixel Mean Power have been isolated in the integrate extraction stage in addition to features like district, boundary, and inconsistency.

The best model stops part extraction and accuracy evaluation after the reveal of the harm handle. However, portraying it as harmless or demeaning has not been done. Similar to this, further analysis of detrimental development control has been carried out using support vector machines. As preparation characteristics, isolated pieces are used, and an organised model is created. Then, utilising that already-prepared conjecture model, a dark recognised dangerous development handle is constructed. 3.2.1 Image Processing When looking at photographs for the first time, the central channel of the grayscale CT image is used. A few controversies are brought up on CT pictures taken at the time of picture taking measurement that helps with fake handle area. Procedia Programming 125 (2018) 107–114 10921, Suren Makaju et al. Show One of the causes of hazard passings is cell disintegration in the lungs. Due to the fact that it appears and exhibits symptoms at a very early stage, it is challenging to identify.

However, early diagnosis and treatment of illness might reduce the passing rate the chance. optimum imaging method

As CT imaging would detect both predicted of undiscovered lung cells breakdown pathways, it's really dependable for this purpose [1]. However, variations in force in CT examination images and actual planning errors by radiologists and trained personnel may cause issues in identifying the [2] Hazardous cell PC Upheld Examination has as of late turned into a reciprocal and promising device to help radiologists and experts in accurately separating the sickness [3].

Various structures have been created, and research is progressing to recognize cell breakdown in the lungs. Nonetheless, a few structures don't have sufficient recognizable proof accuracy, and a few systems really should be upgraded to accomplish the most striking precision drawing nearer 100 percent.

The cell breakdown in the lungs has been recognized and depicted utilizing picture arrangement methods and artificial intelligence processes. To pick the new best structures, we read up continuous systems produced for harm revelation in light of CT check pictures of the lungs. Research was then centered around them, and another model was proposed. Composing Review 2. cell breakdown in the lungs using different approaches of picture planning and artificial intelligence. Aggarwal, Furquan furthermore, Kalra [4] proposed a model that gives portrayal among handles and regular lung life frameworks structure. The procedure eliminates numerical, quantifiable and faint level characteristics. LDA is used as classifier and ideal thresholding for division. The structure has 84% precision, 97.14% affectability and 53.33% disposition. Though the structure recognizes the harm handle, its accuracy is at this point prohibited. No any simulated intelligence techniques has been used to bunch also, essential division techniques is used. As such, blend of any of its means in our new model doesn't give probability of progress. Jin, Zhang and Jin [5] used convolution brain association as classifier in his PC supported plan system to recognize the cell breakdown in the lungs. The system has 84.6% of precision, 82.5% of affectability and 86.7% of identity. The potential gain of this model is that it uses indirect divert in Area of interest (return for capital contributed) extraction stage which reduces the cost of getting ready and affirmation steps. Regardless of the way that, execution cost is decreased, it has still unsatisfactory accuracy. Sangamithraa and Govindaraju [6] uses K mean performance learning estimation for gathering or division. It bundles the pixel dataset according to specific ascribes. For portrayal this model executes back multiplication association. Arrangements like entropy, relationship, homogeneity, PSNR, SSIM are taken out using faint level co-occasion framework (GLCM) method. The framework has accuracy of around 90.7%. Picture pre arranging focus channel is utilized for disturbance clearing which can be valuable for our new model to kill the commotion and work on the exactness. Roy, Sirohi, and Patle [7] urged a construction to see cell breakdown in the lungs handle utilizing comfortable impedance framework and dynamic construction model. This construction involves faint change for picture contrast upgrade. Picture binarization is performed before division and the picture is circulated utilizing the strong design model.\

III. ALGORITHM

CNN A Convolutional Neural Organization (ConvNet/CNN) is a Profound Learning calculation which can take in an information picture, dole out significance (learnable loads and inclinations) to different perspectives/objects . The pre-handling needed in a ConvNet is a lot of lower when contrasted with other characterization calculations. The design of a ConvNet is comparable to that of the availability example of Neurons in the Human Mind and was motivated by the association of the Visual Cortex. Individual neurons react to boosts just in a limited locale of the visual field known as the Open Field. An assortment of such fields cross-over to cover the whole visual region.

Dataset-: We will take a dataset of lung cancer from Kaggle.

PSO The particle swarm optimization (PSO) algorithm is a population-based search algorithm based on the simulation of the social behavior of birds within a flock. The initial intent of the particle swarm concept was to graphically simulate the graceful and unpredictable choreography of a bird flock, to discover patterns that determine a bird's capacity to fly with unison and to quickly change direction by regrouping in the best formation. From this basic goal, the idea developed into a straightforward and effective optimization procedure.

Conclusion

With CNN, our suggested technology will automatically detect lung cancer in photos. Techniques for pre-processing improve the precision of cancer detection. We will apply the CNN classifier, which is a feature of the Weka 30 classifier, to classify the data.

References

[1] Imran Ali, Waseem.A. Wani and Kishore Saleem, “Cancer Scenario in India with Future Perspectives,” Cancer Therapy, 8, pp. 56 - 70, 2011. [2] Beadsmoore CJ, Screaton NJ, “Classification, staging and prognosis of lung cancer,” European Journal of Radiology, 4(1), pp. 8 - 17, 2003. [3] Todd M.Blodgett, Ajeet S. Mehta, Amar S.Mehta, Charles M. Laymon, Jonathan Carney, David W. Townsend, “PET/CT artifacts” Clinical Imaging, 35, pp. 49 - 63, 2011. [4] Robert M Haralick, K.Shanmugam and Its’hak Dinstein, “Textural Features for Image Classification,” IEEE Transactions on Systems, Man and Cybernetics, 3(6), pp. 610 - 621, 1973. [5] S.A. Patil, M.B. Kuchanur, “Lung Cancer Classification using Image Processing,” International Journal of Engineering and Innovative Technology (IJEIT), 2(3), pp. 37 – 42, 2012. [6] AjmedS.Al-Fehoum, Eslam B.Jaber, Mohammed A.Al-Jarrah “Automated detection of lung cancer using statistical and morphological image processing techniques,” Journal of Biomedical Graphics and Computing, 4(2), pp. 33 - 42, 2014. [7] Gang Xie, Tianrui Cao, Chengdong Yan, “Texture Features Extraction of Chest HRCT Images based on Granualr Computing,” Journal of Multimedia, 5(6), pp. 639 – 647, 2010. [8] Qi Song, Junjie Bai, Dongfeng Han, Sudershan Bhatia, Wenqing Sun, William Rockey, John E. Bayouth, John M.Buatti and Xiaodong Wu, “Optimal Co-Segmentation of Tumor in PET-CT Images with Context Information,” IEEE Transactions on Medical Imaging, 32(9), pp. 1685 – 1697, 2013. [9] C. Panyindee and W. Chiracharit, “ Detection of Lung Tumors in Chest PET/CT Images by using Watershed Transform and Image Mapping,” The 3rd International Symposium on Biomedical Engineering (ISBME 2008), pp. 9 – 12, 2008. [10] K.A.G.Udeshani, R.G.N. Meegama, T.G.I.Fernando, “Statistical Feature - based Neural Network Approach for the Detection of Lung Cancer in Chest X-Ray Images,” International Journal of Image Processing (IJIP), 5(4), pp. 425 – 434, 2011 [11] G.S. Cox, F.J. Hoare, G.de Jager, “Experiments in Lung cancer nodule detection using texture analysis and neural network classifiers,” Third South African Workshop on Pattern Recognition, Pretoria 1992. [12] M.S. Ahmad, M. Shahid Naweed and M. Nisha “Application of Texture Analysis in the Assessment of Chest Radiographs,” International Journal of Video & Image Processing and Network Security IJVIPNS-IJENS, 9(9), pp. 32 – 36, 2009. [13] Balaji Ganeshan, Sandra Abaleke, Rupert C.D. Young, Christopher R Chatwin and Kenneth A. Miles, “Texture analysis of non-small cell lung cancer on unenhanced computed tomography: initial evidence for a relationship with tumour glucose metabolism and stage,” Cancer Imaging, 10(1), pp. 137-143, 2010. [14] Ada, Rajneet Kaur, “Early Detection and Prediction of Lung Cancer Survival using Neural Network Classifier,” International Journal of Application or Innovation in E ngineering & Management, 2(6), pp. 375 – 383, 2013. [15] Mir Rayat Imtiaz Hossain, Imran Ahmed, Md. Hasanul Kabir, “Automatic Lung Cancer Detection using GLCM features,” Asian Conference on Computer Vision ACCV’14, Singapore, 2014 [16] Guo Xiuhua, Sun Tao, Wu Haifeng, He Wen, Liang Zhigang, Zhang Mengxia, GuoAimin, Wang Wei, “Support Vector Machine Prediction Model of Early – stage Lung Cancer Based on Curvelet Transform to Extract Texture Features of CT Image,” World Academy of Science, Engineering and Technology, 4(11), pp. 278 – 282, 2010. [17] Sandeep A. Dwivedi, R.P. Borse, Anil M.Yametkar, “ Lung Cancer detection and Classification by using Machine Learning & Multinomial Bayesian,” IOSR Journal of Electronics and Communication Engineering (IOSR – JECE), 9(1), pp. 69 – 75, 2014. [18] Vinod Kumar, Kanwal Garg, Vijay Kher, “Early Diagnosis of Lung Cancer with ANN, FCM and FMNN,” International Journal of Advanced Research in Computer Science and Software Engineering,

Copyright

Copyright © 2023 Prof. G. S Mate, Karan Shirsat, Aakash Panditha, Kaumodaki Koul, Sejal Rayou. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.