Search or add a thesis

Advanced Search (Beta)
Home > Segmentationandclassificationofbraintumormrimages Throughadvancemachinelearning Methods

Segmentationandclassificationofbraintumormrimages Throughadvancemachinelearning Methods

Thesis Info

Access Option

External Link

Author

Farhi, Lubna

Program

PhD

Institute

National University of Sciences & Technology

City

Islamabad

Province

Islamabad

Country

Pakistan

Thesis Completing Year

2018

Thesis Completion Status

Completed

Subject

Electrical Engineering

Language

English

Link

http://prr.hec.gov.pk/jspui/bitstream/123456789/9893/1/Lubna%20Farhi_EE_2018_NUST_Main%20part.pdf

Added

2021-02-17 19:49:13

Modified

2024-03-24 20:25:49

ARI ID

1676727830905

Similar


The main objective of this research work is to develop, test and evaluate an identification support system that is able to provide accurate, fast and reliable diagnosis of brain tumor in MRimages. Keeping in consideration that human decision making skills are mainly dependent on experience and prone to error due to fatigue, Artificial Intelligence (AI) can be utilized as an effective aid in the field of medicinal sciences for tumor diagnosis through image recognition. Therefore, this thesis strives to develop such an intelligent system that can be used for the segmentation and classification of infiltrative brain tumors known as Low Grade and High Grade in MR images. In order to tackle the complex task of brain tumor segmentation in MR images, we present an adaptive algorithm that formulates an energy based stochastic segmentation with a level set methodology. This hybrid technique efficiently matches, segments and determines the anatomic structures within an image by using global and local energies. After evaluating the algorithm on low and high grade images, it was noted that there was an improvement in the resultant similarity between segmented and truth (original) images. Once effective segmentation was achieved we could then work on the next step of tumor identification; classification. In the second part of the process we proposed two classification frameworks, machine learning and deep learning. In machine learning, we first extracted 22 probabilistic features using gray level co-occurrence matrix methodology that served as input features for the classifiers. Then we showed the improvement in classification (through machine learning) accuracy by providing two methodologies in which the first one involved v classification directly after feature extraction whereas in the second we reduced the extracted features using principal component analysis and then applied those reduced features to several classifiers. The second framework that we proposed was the brain tumor classification of segmented MR images through optimized CNN-Deep belief learning model. It scales to various image sizes by distributing the hyper-parameters and weights among all locations in an image. The presented model is translation invariant and is compatible with top-down and bottom-up probabilistic inference. This hierarchical classifier was optimized by regularization, that mitigates the effect of overfitting for small datasets, stochastic gradient decent, which works efficiently by utilizing only a small set of samples from a whole training set to infer the gradient and fine tuning of constraints. A comparative analysis, based on accuracy, error/loss and computation time, was carried out between the pre-processed non-segmented and segmented MR images after classification was completed. The results showed that the accuracy of proposed optimized CNN-deep belief learning classifier with segmented MR images was higher while the loss and execution time were reduced. These methodologies transcend the confines of MR image processing due to their effective modularity allowing them to be suitable for other medical imaging and computer vision tasks.
Loading...
Loading...

Similar Books

Loading...

Similar Chapters

Loading...

Similar News

Loading...

Similar Articles

Loading...

Similar Article Headings

Loading...

استغاثہ بحضور سرورِ کائناتﷺ

استغاثہ
بحضور سرورِ کائناتؐ

جہل و ظلمت ہیں گھیرے ہمیں چار سو
ہم کو درکار ہے روشنی یانبیؐ

حشر میں اک سماں ہو گا دیکھیں گے جب
چہرۂ مصطفیٰؐ اُمتی یانبیؐ

نور ہی نور ہیں احمدؐ و فاطمہؑ
ہوں حسینؑ و حسنؑ کہ علیؑ یانبیؐ

ہو نگہ اک، بھنور میں ہے امت گھری
ہم کو گھیرے ہے اب تیرگی یانبیؐ

آپ کی رحمتوں میں زمین و زمن
آپ ہی سے ملی زندگی یانبیؐ

حق ہوا جلوہ گر آپ کی ذات میں
حق کی ہیں آپؐ ہی روشنی یانبیؐ

مجھ فضاؔ کے لیے ہے یہ سامانِ حشر
نعت میں نے جو یہ ہے لکھی یانبیؐ

A صاحبِ تفسیر ِصدیقی کا عصری مسائل کے حل اور کلامی مباحث میں طرزِاستدلال: تجزیاتی مطالعہ Sahib-e-Tafseer-e-Sadiqi's style of reasoning in contemporary problem solving and discourse discussions: an analytical study

Holy Qur’an is the last revealed book of Allah Almighty.The explanation of its verses started from the time of last apostle and it will continue till the day of resurrection. The land of Indo Pak sub-continent is much fertile regarding the personalities who worked for the interpretation of the last revealed book. In this research paper characteristics and methodology of Tafseer-e-Siddiqui are discussed, especially Theological Discussions of exegesis are analysed. Moulana Abdul Qadeer Siddiqui was a renowned theologian of Hyderabad Dakkan. He spent his whole life in learning and teaching Islam. His work on Tafseer-e-Qur’an is a great contribution for Quranic understanding. In this Tafseer the writer has also consulted books of other religions.He criticized orientalists but with politeness. There is dire need to spread this contribution of Moulana Siddiqui among the Muslims and especially the students of educational institutions.                                                                                              Key Words: Holy Qur’an, Hadith, Orientialists, Chiristianity, Judaism.            

Preparation of High Specific Activity Radionuclides for Therapeutic Radiopharmaceuticals

Production of radioactive scandium by irradiating natural titanium metal in Pakistan Research Reactor-1 was evaluated. The production rate of scandium-47 ( 47 Sc) and other radioactive scandium was estimated. High specific activity 47 Sc can be produced by irradiating enriched titanium-47 in sufficient quantities needed for therapeutic applications. A new separation technique based on column chromatography was developed. Neutron irradiated titanium was dissolved in hydrofluoric acid, which was evaporated and taken in distilled water. The resulting solution was loaded on silica gel column. The radioactive scandium comes out first and the inactive titanium is removed with 2 M HCl. More than 95% radioactive scandium was recovered, while chemical impurity of titanium determined by optical emission spectroscopy was less than 0.01 μg / mL in final product of 47 Sc. Production of Copper-64 ( 64 Cu) by irradiating copper and zinc metals in a reactor was evaluated. Low specific activity 64 Cu can be easily produced using thermal neutrons via 63 Cu (n, γ) 64 Cu reaction, while use of fast neutrons are mandatory for high specific activity 64 Cu via 64 Zn (n, p) 64 Cu reaction. Natural copper and zinc targets were irradiated in Pakistan Research Reactor-1. Radionuclidic impurities produced by thermal and fast neutrons were determined. Commonly available organic anion exchange resin (AG 1-X8) was used for the separation of no-carrier-added radiocopper from neutron irradiated zinc. More than 95 % 64,67 Cu was recovered. The radionuclidic and chemical purity of 64 Cu was determined. The specific activity of 64 Cu produced by 63 Cu (n, γ) and 64 Zn (n, p) was compared.The metallic cation, 68 Ga (III) is suitable for complexation with chelators either naked or conjugated with biological macromolecules, however, such labeling procedure requires high chemical purity and concentrated solutions of 68 Ga (III), which cannot be sufficiently fulfilled by the presently available 68 Ge/ 68 Ga generator eluate. A method to increase the concentration and purity of generator has been developed. The 68 Ga obtained from a commercial 68 Ge/ 68 Ga 68 Ga eluate (1M HCl) is extracted in methyl ethyl ketone, which is evaporated and taken in a small volume of buffer. Arsenic-77 (T 1/2 = 1.6 d) was produced by irradiating natural germanium in Pakistan Research Reactor-1. The nuclear reaction 76 Ge (n, γ) produces 77 Ge, which decays by emission of β - particles into 77 As. The neutron irradiated target was dissolved in aqua regia, excess of acid was removed by evaporation and finally the solution in basic media was passed through hydrous zirconium oxide (HZO) column. The Ge was quantitatively retained on HZO, while 77 As was present in the effluent. More than 90 % 77 As was recovered. The chemical impurity of Ge in 77 As was <0.01μg/mL. Large columns containing aluminum oxide (Al 2 O 3 ) or gel (e.g. zirconium molybdate) are needed to prepare 98 Mo(n,γ) 99 Mo→ 99m Tc column chromatographic generators that results in large elution volumes containing relatively high 99 Mo impurity and low concentrations of 99m Tc. Post elution concentration of 99m Tc using in house prepared lead cation exchange and alumina column was developed. The principle of the method developed is trapping of anionic pertechnetate on tiny alumina column. This can be only achieved in the absence of sulfate ions. These sulfate ions are removed from the eluate by reaction with lead ions loaded onto a cation exchange column, to precipitate lead sulfate, which is filtered out by the column packing. Using these columns high bolusvolumes (10-60 ml 0.02 M sodium sulfate) of 99m Tc can conveniently be concentrated in 1 mL of physiological saline. This approach also works very effectively to prepare high specific volume solutions of 99m Tc-pertechnetate from a fission based 99 Mo/ 99m Tc generator in the second week of its normal working life. Rhenium-188 is also obtained from alumina based 188 W→ 188 Re generator, and developed technique can also be used for the concentration of 188 Re. Because of the high content of inactive molybdenum in neutron irradiated MoO 3 , large columns containing alumina or gel are needed to produce chromatographic 99 Mo→ 99m Tc generator. This results in large elution volumes containing relatively high 99 Mo breakthrough and low concentrations of of 99m 99m TcO 4- . The decrease in specific volume Tc places a limitation on reconstitution of some kits for 99m Tc radiopharmaceuticals applied in diagnostic nuclear medicine. Hence concentration technique is mandatory for effective utilization of (n,γ) produced 99 Mo/ 99m Tc generators at the start of its life whereas in case of fission 99 Mo/ 99m Tc generator the technique may be quite useful at the end of first week of its life. Post elution concentration of 99m Tc using in house prepared lead (Pb) column was developed. The high bolus volumes (10-60 ml saline) can conveniently be concentrated in ~1 ml of saline. The adsorption behavior of Na 188 ReO 4 is quite different from Na 99m TcO 4 on lead column. Sodium perrhenate did not adsorb on Lead column and found quantitatively in effluent. Thus Lead column may also be used for the separation of Na 99m TcO 4 from Na 188 ReO 4 . The high bolus volumes (20–40 mL) of the generator-produced Rhenium-188 require post elution concentration of the eluate for the preparation of a dissolved β − source and radiopharmaceuticals labeled with Re-188 for radiotherapy. Solvent extractionof 188 Re in methyl ethyl ketone was studied. With the increase of organic phase volume, extraction of 188 Re was enhanced while mixing time of aqueous and organic phases did not show any significant effect on the extractability of 80% of 188 Re in the organic phase. Almost 188 Re was extracted in methyl ethyl ketone at a volume ratio of 1 : 2 for aqueous and organic phases. By evaporation/distillation of methyl ethyl ketone, concentrated and dissolved in the desired volume of physiological saline.