A dissertation submitted to the Faculty of Health Sciences, University of the Witwatersrand, in the fulfillment of the requirements for the degree of Master of Medicine in Ophthalmology. Johannesburg, February 2017 / Purpose: To evaluate if corneal collagen crosslinking carried out on patients with keratoconus, slows down or halts the progression of keratoconus. To determine which group of keratoconus patients benefited most from the procedure. Methods: A retrospective record review of 41 eyes of 29 patients. Visual acuity and keratometry measurements were recorded for the involved eye pre-crosslinking and at 3 months and 6 months post-crosslinking. A comparison of these variables pre-crosslinking and at 6 months post-crosslinking was made to determine if there was a flattening of corneal curvature (keratometry readings) and an improvement in visual acuity. Patients were further divided into 3 groups of keratoconus, based on their keratometry readings (measured in diopters): mild keratoconus (≤47 diopters), moderate keratoconus (48 – 54 diopters) and advanced keratoconus (≥55 diopters), to determine which group of keratoconus had the best keratometry reduction readings. Results: After crosslinking took place on 41 eyes, the UnVA of 16(39%) eyes showed an improvement at 6 months, 17(41%) eyes showed no change and 8(20%) eyes showed a decrease in UnVA at 6 months, compared to pre-CXL values. For BCVA, 12(29%) eyes showed an improvement at 6 months, 18(44%) eyes showed no change and 11(27%) eyes showed a decrease in BCVA at 6 months, compared to pre-CXL values. Keratometry readings however showed that 23(56%) eyes had an average flattening of corneal curvature readings of 0.7 D and the remaining 18(44%) eyes showed more steepening (worsening) of the corneal curvature readings of 0.9 D after 6 months post-CXL. 30(73%) eyes had mild keratoconus, 7(17%) had moderate keratoconus and 4(10%) had advanced keratoconus. 19 of the 30 eyes in the mild keratoconus group (73%) showed an average flattening of corneal curvature of 0.6 D. 4 of the 7 eyes in the moderate keratoconus group (17%) showed an average flattening of corneal curvature of 0.7 D. All 4 patients in the advanced group (10%) had steepening (worsening) of their corneal curvatures with an average of 1.2 D. Conclusion: Corneal collagen crosslinking performed on keratoconus patients at least halts the progress of keratoconus. 6 months after CXL most patients showed minimal change from pre-CXL to 6 months in both visual acuity and keratometry. However a longer follow up period and larger sample size is needed to determine if vision and keratometry readings can improve significantly. / MT2017

Background: Today corneal crosslinking with ultraviolet-A photoactivation of riboflavin is an established method to halt the progression of keratoconus. In some cases, when the refractive errors are large and the visual acuity is low, conventional corneal crosslinking may not be sufficient. In these cases it would be desirable with a treatment that both halts the progression and also reduces the refractive errors and improves the quality of vision. Aims:  The aims of this thesis were to determine whether mechanical compression of the cornea during corneal crosslinking for keratoconus using a sutured rigid contact lens could improve the optical and visual outcomes of the treatment, and also to find methods to evaluate the effect of different corneal crosslinking treatment regimens. Methods: In a prospective, open, randomized case-control study, 60 eyes of 43 patients with progressive keratoconus, aged 18-28 years, planned for routine corneal crosslinking, and a corresponding age- and sex-matched control group was included. The patients were randomized to conventional corneal crosslinking (CXL; n=30) or corneal crosslinking with mechanical compression of the cornea during the treatment (CRXL; n=30). Biomicroscopy, autorefractometry, best spectacle corrected visual acuity, axial length measurement, Pentacam® HR Scheimpflug photography, pachymetry, intraocular pressure measurements and corneal biomechanical assessments were performed before treatment (baseline) and at 1 month and 6 months after the treatment. One of the articles evaluated and compared the optical and visual outcomes between CXL and CRXL, while the other three articles focused on methods to evaluate treatment effects. In Paper I, the corneal light scattering was manually quantified from Scheimpflug images throughout the corneal thickness at 8 measurements points, 0.0 to 3.0 mm from the corneal centre, in patients treated with CXL. In Paper IV the corneal densitometry (light scattering) was measured with the Pentacam® HR software, in 4 circular zones around the corneal apex and at 3 different depths of the corneal stroma, in both CXL and CRXL treated corneas. Paper III quantified the biomechanical effects of CXL in vivo. Results: Corneal light scattering after CXL showed distinctive spatial and temporal profiles and Applanation Resonance Tonometry (ART) -technology demonstrated an increased corneal hysteresis 1 and 6 months after CXL. When comparing the refractive and structural results after CXL and CRXL, CRXL failed to flatten the cornea, and the treatment did not show any benefits to conventional CXL treatment, some variables even indicated an inferior effect. Accordingly, the increase in corneal densitometry was also less pronounced after CRXL. Conclusions: Analysis of corneal light scattering/densitometry shows tissue changes at the expected treatment location, and may be a relevant variable in evaluating the crosslinking effect. ART -technology is an in vivo method with the potential to assess the increased corneal hysteresis after CXL treatment. By refining the method, ARTmay become a useful tool in the future. Unfortunately, CRXL does not improve the optical and visual outcomes after corneal crosslinking. Possibly, stronger crosslinking would be necessary to stabilize the cornea in a flattened position.

<p>Clinical imaging within ophthalmology has had transformative effects on ocular health over the last century. Imaging has guided clinicians in their pharmaceutical and surgical treatments of macular degeneration, glaucoma, cataracts and numerous other pathologies. Many of the imaging techniques currently used are photography based and are limited to imaging the surface of ocular structures. This limitation forces clinicians to make assumptions about the underlying tissue which may reduce the efficacy of their diagnoses. </p><p>Optical coherence tomography (OCT) is a non-invasive, non-ionizing imaging modality that has been widely adopted within the field of ophthalmology in the last 15 years. As an optical imaging technique, OCT utilizes low-coherence interferometry to produce micron-scale three-dimensional datasets of a tissue's structure. Much of the human body consists of tissues that significantly scatter and attenuate optical signals limiting the imaging depth of OCT in those tissues to only 1-2mm. However, the ocular anterior segment is unique among human tissue in that it is primarily transparent or translucent. This allows for relatively deep imaging of tissue structure with OCT and is no longer limited by the optical scattering properties of the tissue. </p><p>This goal of this work is to develop methods utilizing OCT that offer the potential to reduce the assumptions made by clinicians in their evaluations of their patients' ocular anterior segments. We achieved this by first developing a method to reduce the effects of patient motion during OCT volume acquisitions allowing for accurate, three dimensional measurements of corneal shape. Having accurate corneal shape measurements then allowed us to determine corneal spherical and astigmatic refractive contribution in a given individual. This was then validated in a clinical study that showed OCT better measured refractive change due to surgery than other clinical devices. Additionally, a method was developed to combine the clinical evaluation of the iridocorneal angle through gonioscopy with OCT.</p> / Dissertation

Toric intraocular lenses (IOLs) are designed to reduce spectacle dependency by correcting corneal astigmatism at the time of surgery. However, these IOLs are reliant on the accurate prediction of post-operative corneal astigmatism through reliable ocular biometry and the accurate calculation of surgically induced astigmatism. In the thesis the repeatability of assessing corneal curvature was assessed using six commercially available keratometers. The results question the validity of corneal biometry and infer that much of the apparent change in corneal shape usually associated with surgically induced astigmatism may be due to measurement error. The use of the oblique cross cylinder formulae for the calculation of post-operative corneal curvature was also investigated. This formula is incorporated into all commercially available toric IOL calculators and is utilised in every toric IOL implantation. The results from this thesis indicate that the formula is not applicable to the human cornea and that the use of the calculator does not increase the effectivity of the toric correction. Furthermore, the thesis queries the assumption that post-operative corneal astigmatism is directly proportional to post-operative refractive error. The disparity between both the magnitude and axis of astigmatism measured by keratometry and manifest refraction in a pseudophakic population was investigated. The axis measurements in particular showed very poor agreement; far outside an acceptable level of misalignment, significantly decreasing the effective correction provided if the lens was aligned with the keratometry readings. Inclusion of the posterior corneal curvature and thickness, along with a smaller chord length may lead to a more accurate assessment of corneal power. Despite the difficulty in providing an effective toric IOL correction, it was found that the correction of corneal astigmatism at the time of cataract surgery might decrease the risks of falls. Uncorrected astigmatism and cataract both cause a reduction in stability when stepping oven an obstacle, which is one of the most common causes of trips and falls in the elderly population.

Neste trabalho e apresentado um software baseado em técnicas de processamento digital de imagens para medida de astigmatismo corneano para dispositivos portáteis. O sistema físico, um dispositivo adaptado para lâmpadas de fenda, projeta 36 pontos luminosos dispostos em um formato preciso de circulo na região central da córnea. O deslocamento, o tamanho e a deformação da imagem refletida destes pontos luminosos pela córnea são analisados proporcionando a ceratometria. O software apresentado neste projeto foi desenvolvido para poder ser instalado nos principais dispositivos moveis que disponibilizem a plataforma Java ME, ou seja, pode ser implementado em computadores portáteis sem depender da plataforma utilizada. Este projeto permitiu a viabilidade de portabilidade para o ceratômetro, ou seja, imagens originais do ceratômetro em Lâmpada de Fenda, desenvolvido pelo Laboratório de Instrumentação Oftálmica, foram inseridas em dispositivos móveis e processadas. As imagens em formato BMP, de tamanho típico de 640x480 pixels são processadas e os 36 pontos luminosos são detectados em media em 1150 ms. Então, e traçada a elipse que melhor se ajusta a estes pontos e os raios de maior e menor dimensão são calculados em pixels. Todo o processo e realizado em media em 1199 ms. A contribuição deste trabalho e mostrar a viabilidade de transformar o ceratômetro de lâmpada de fenda em um aparelho portátil. / This work is about software development based on digital image processing techniques for corneal astigmatism measurements for portable devices. The physical system comprises a device adapted for slit lamps and projects 36 light spots displayed in a precise circle at the lachrymal film of the examined cornea. Keratometry is provided by the analisys of displacement, size and deformation of the reflected image of the light spots. The software presented in this project was developed to be installed mainly for mobile devices that provide the Java ME platform, ie, it can be implemented in portable computers. Furthermore, it allows portability for keratometers and is a previous work for a portable corneal topographer It also shows viability of portability for the keratometer, ie, the original keratometer images in Slit Lamp, developed by the Laboratory of Ophthalmic Instrumentation, were embedded in mobile devices and have been processed. The BMP type images, - typical size of 640x480 pixels - are processed and the 36 points of light are detected in an average of 1150 ms. The best fitting ellipse in the reflected image of these points provides the major and minor elpise radii in pixels. The entire process is performed in 1199 ms as an avarage. The contribution of this work is to show the viaibility of transforming the keratometer for Slit Lamps in a handheld device.

Neste trabalho e apresentado um software baseado em técnicas de processamento digital de imagens para medida de astigmatismo corneano para dispositivos portáteis. O sistema físico, um dispositivo adaptado para lâmpadas de fenda, projeta 36 pontos luminosos dispostos em um formato preciso de circulo na região central da córnea. O deslocamento, o tamanho e a deformação da imagem refletida destes pontos luminosos pela córnea são analisados proporcionando a ceratometria. O software apresentado neste projeto foi desenvolvido para poder ser instalado nos principais dispositivos moveis que disponibilizem a plataforma Java ME, ou seja, pode ser implementado em computadores portáteis sem depender da plataforma utilizada. Este projeto permitiu a viabilidade de portabilidade para o ceratômetro, ou seja, imagens originais do ceratômetro em Lâmpada de Fenda, desenvolvido pelo Laboratório de Instrumentação Oftálmica, foram inseridas em dispositivos móveis e processadas. As imagens em formato BMP, de tamanho típico de 640x480 pixels são processadas e os 36 pontos luminosos são detectados em media em 1150 ms. Então, e traçada a elipse que melhor se ajusta a estes pontos e os raios de maior e menor dimensão são calculados em pixels. Todo o processo e realizado em media em 1199 ms. A contribuição deste trabalho e mostrar a viabilidade de transformar o ceratômetro de lâmpada de fenda em um aparelho portátil. / This work is about software development based on digital image processing techniques for corneal astigmatism measurements for portable devices. The physical system comprises a device adapted for slit lamps and projects 36 light spots displayed in a precise circle at the lachrymal film of the examined cornea. Keratometry is provided by the analisys of displacement, size and deformation of the reflected image of the light spots. The software presented in this project was developed to be installed mainly for mobile devices that provide the Java ME platform, ie, it can be implemented in portable computers. Furthermore, it allows portability for keratometers and is a previous work for a portable corneal topographer It also shows viability of portability for the keratometer, ie, the original keratometer images in Slit Lamp, developed by the Laboratory of Ophthalmic Instrumentation, were embedded in mobile devices and have been processed. The BMP type images, - typical size of 640x480 pixels - are processed and the 36 points of light are detected in an average of 1150 ms. The best fitting ellipse in the reflected image of these points provides the major and minor elpise radii in pixels. The entire process is performed in 1199 ms as an avarage. The contribution of this work is to show the viaibility of transforming the keratometer for Slit Lamps in a handheld device.

Submitted by RODOLFO DE LIMA FACHINELLI null (r_fachinelli@hotmail.com) on 2018-04-04T18:54:51Z No. of bitstreams: 1 DISSERTAÇÃO - RODOLFO FACHINELLI.pdf: 1066915 bytes, checksum: 93987ac1254ae543ed84ff519a438005 (MD5) / Approved for entry into archive by ROSANGELA APARECIDA LOBO null (rosangelalobo@btu.unesp.br) on 2018-04-06T13:56:06Z (GMT) No. of bitstreams: 1 fachinelli_rl_me_bot.pdf: 1066915 bytes, checksum: 93987ac1254ae543ed84ff519a438005 (MD5) / Made available in DSpace on 2018-04-06T13:56:06Z (GMT). No. of bitstreams: 1 fachinelli_rl_me_bot.pdf: 1066915 bytes, checksum: 93987ac1254ae543ed84ff519a438005 (MD5) Previous issue date: 2018-02-22 / Objetivo: Comparar a ceratometria média aferida (KA) sob narcose em crianças submetidas à facectomia, por catarata congênita ou do desenvolvimento, com a ceratometria média obtida por cálculo teórico (KC), utilizando o poder dióptrico da lente intraocular (LIO) implantada e dados refracionais e biométricos pós-operatórios, a fim de evidenciar possíveis erros de aferição da KA devido a narcose. Métodos: Estudo retrospectivo realizado a partir da análise de dados coletados de prontuários de pacientes com catarata bilateral, congênita ou do desenvolvimento, que receberam tratamento cirúrgico no HC-FMB. Foram analisados 73 olhos de crianças que possuíam pelo menos um exame pós-operatório completo, cada momento em que determinado olho foi examinado foi considerado um elemento do conjunto amostral, totalizando 165 momentos. KC foi obtida a partir de fórmula teórica para cálculo de LIO, utilizando os dados biométricos pós-operatórios (AL e ACD), refratometria automatizada pós-operatória e poder dióptrico da LIO implantada. Foi calculada a diferença entre KA e KC (Desvio = KA – KC). Para análise estatística dos desvios encontrados e comparação entre KA e KC, foi realizado o teste de associação de Goodman e o teste não paramétrico de Kruskal-Wallis. Resultados: A média de idade no momento da cirurgia do primeiro olho foi 954,62 dias, com desvio padrão de ±794,14 dias, mediana de 953 dias, sendo a idade mínima de 44 dias e a máxima de 2659 dias. Vinte e três eram do sexo masculino (62,16%) e 14 eram do sexo feminino (37,84%). KA variou de 40,62D a 51,50D, com mediana de 45,25D, média de 45,32D e desvio padrão de ±2,37D. KC variou de 39,40D a 52,26D, com mediana de 44,49D, média de 44,54D e desvio padrão de ±2,41D. Os desvios (Desvio = KA – KC) variaram de -2,28D a 3,81D, com mediana de 0,83D, média de 0,79D e desvio padrão de ±1,18D. A relação entre KC e KA pode ser representada pela equação KA = 1,0172 Kc. Conclusão: A comparação entre KC e KA em crianças sob narcose evidenciou que há superestimação do valor aferido em relação ao calculado. A análise dos desvios encontrados mostrou tendência para maior superestimação quanto maior a KA com diferença significativa (p<0,05) nas aferições acima de 44,0D. / Purpose: To compare the mean keratometry measured (KA) in children under anesthesia to receive surgical treatment for congenital or developmental cataract with a mean keratometry obtaeined by a theoretical formula (KC), using the implanted intraocular lens (IOL) power value and postoperative refractional and biometric data, in order to check possible observational erros of KA due to general anesthesia. Methods: A retrospective study analysing records of patients with congenital or developmental bilateral cataract who underwent surgical treatment at HC-FMB. Seventy-three children’s eyes that had at least one full postoperative exam were analyzed, each time one eye was assessed was considered one element of the sample group, and the total amount was 165 elements. KC was determined by a theoretical formula for calculating IOL power using postoperative biometric data (AL e ACD), postoperative automatic refractometery and refractive power of the implanted IOL. The KA observational error value was obtained by subtracting KC from KA (Error = KA - KC). Statistical analysis of the observational erros and the comparison between KA and KC were made by Goodman’s test and Kruskal-Wallis’ non-parametric test. Results: the mean age at the moment of the first surgery was of 954,62 days, the standard deviation was ±794,14 days, the median was 953 days, the minimum age was 44 days and the maximum age was 2659 days. Twenty-three (62,16%) patients were male and 14 (37,84%) female. KA ranged from 40,62D to 51,50D, the median was 45,25D, the mean was 45,32D and the standard deviation was ±2,37D. KC ranged from 39,40D to 52,26D, the median was 44,49D, the mean was 44,54D and the standard deviation was ±2,41D. Observational errors ranged from -2,28D to 3,81D, the median was 0,83D, the mean was 0,79D and the standard deviation was ±1,18D. The rate between KC and KA can be represented by the equation KA = 1.0172 KC. Conclusion: The comparison between KC and KA in children under general anesthesia showed that there is an overestimation of the value measured when compared to the calculated one. The analysis of the observational errors showed there is a tendency to greater overestimation the higher the KA, with significant difference (p <0.05) in the measurements over 44.0D.

Orientador: Antonio Carlos Lottelli Rodrigues / Resumo: Objetivo: Comparar a ceratometria média aferida (KA) sob narcose em crianças submetidas à facectomia, por catarata congênita ou do desenvolvimento, com a ceratometria média obtida por cálculo teórico (KC), utilizando o poder dióptrico da lente intraocular (LIO) implantada e dados refracionais e biométricos pós-operatórios, a fim de evidenciar possíveis erros de aferição da KA devido a narcose. Métodos: Estudo retrospectivo realizado a partir da análise de dados coletados de prontuários de pacientes com catarata bilateral, congênita ou do desenvolvimento, que receberam tratamento cirúrgico no HC-FMB. Foram analisados 73 olhos de crianças que possuíam pelo menos um exame pós-operatório completo, cada momento em que determinado olho foi examinado foi considerado um elemento do conjunto amostral, totalizando 165 momentos. KC foi obtida a partir de fórmula teórica para cálculo de LIO, utilizando os dados biométricos pós-operatórios (AL e ACD), refratometria automatizada pós-operatória e poder dióptrico da LIO implantada. Foi calculada a diferença entre KA e KC (Desvio = KA – KC). Para análise estatística dos desvios encontrados e comparação entre KA e KC, foi realizado o teste de associação de Goodman e o teste não paramétrico de Kruskal-Wallis. Resultados: A média de idade no momento da cirurgia do primeiro olho foi 954,62 dias, com desvio padrão de ±794,14 dias, mediana de 953 dias, sendo a idade mínima de 44 dias e a máxima de 2659 dias. Vinte e três eram do sexo masculino (62,... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: Purpose: To compare the mean keratometry measured (KA) in children under anesthesia to receive surgical treatment for congenital or developmental cataract with a mean keratometry obtaeined by a theoretical formula (KC), using the implanted intraocular lens (IOL) power value and postoperative refractional and biometric data, in order to check possible observational erros of KA due to general anesthesia. Methods: A retrospective study analysing records of patients with congenital or developmental bilateral cataract who underwent surgical treatment at HC-FMB. Seventy-three children’s eyes that had at least one full postoperative exam were analyzed, each time one eye was assessed was considered one element of the sample group, and the total amount was 165 elements. KC was determined by a theoretical formula for calculating IOL power using postoperative biometric data (AL e ACD), postoperative automatic refractometery and refractive power of the implanted IOL. The KA observational error value was obtained by subtracting KC from KA (Error = KA - KC). Statistical analysis of the observational erros and the comparison between KA and KC were made by Goodman’s test and Kruskal-Wallis’ non-parametric test. Results: the mean age at the moment of the first surgery was of 954,62 days, the standard deviation was ±794,14 days, the median was 953 days, the minimum age was 44 days and the maximum age was 2659 days. Twenty-three (62,16%) patients were male and 14 (37,84%) female. KA ranged from... (Complete abstract click electronic access below) / Mestre

This work aims to improve the predictability of asphericity and average keratometry in keratoconus patients after implantation of intrastromal corneal ring segments (ICRS) by creating computational models based on machine learning, using tomographic data of the cornea. This study included 209 eyes of 160 keratoconus (grades I, II and III) implanted with ICRS. The Ferrara ICRS with 160 degrees of arch was implanted in all patients. The ICRS thickness varied from 150 to 250 micra. One or two segments were implanted. The base was composed of corneal tomography Pentacam® (Oculus, Wetzlar, Alemanha) parameters, clinical data and Ferrara ring data totaling 39 parameters. To create the models, neural network algorithms type multlayer perceptron (MLP) and linear regression were used. This study was conducted in four phases: (1) Preparation of the database and setting the values to be predicted mean keratometry and asphericity; (2) Calculation of the variation mean keratometry and asphericity and the nomogram calculation error; (3) Application of machine learning algorithms and attribute selection; (4) Mean keratometry and asphericity variation calculation provided for comparing algorithm with the variation of the preoperative and postoperative calculation of the algorithm and of the error. As a result, the best mean absolute error value found for asphericity was 0.19 and mean keratometry was 1.18. Comparing the mean absolute error values of the nomogram and the average absolute error of the algorithm, there was an improvement of 0.11 to asphericity and 0.09 to mean keratometry in relation to the current nomogram, confirming that the use of computational models can achieve more accurate results may contribute to surgical decision in an attempt to improve the quality of vision of keratoconus patients. / O presente trabalho destina-se a melhorar a previsibilidade da asfericidade e da ceratometria média no pós-operatório de implante de anel intraestromal (SAIC) em pacientes com ceratocone por meio da criação de modelos computacionais baseados em aprendizagem de máquina, utilizando dados tomográficos da córnea. Foram incluídos 209 olhos de 160 pacientes com graus I, II e III de ceratocone submetidos a cirurgia com implante de SAIC. Em todos os pacientes foi implantado anel de Ferrara com 160o de arco com variação de espessura entre 150 e 250 µm e presença de 1 ou 2 segmentos. A base foi composta por parâmetros da tomografia de córnea pentacam®, dados clínicos e dados do anel de Ferrara totalizando 39 parâmetros. Para criação dos modelos, foram utilizados os algoritmos Rede Neural do tipo multlayer perceptron (MLP) e regressão linear. Este estudo foi desenvolvido em 4 fases distintas: (1) Preparação da base de dados e definição dos valores a serem preditos de ceratometria média e asfericidade; (2) Cálculo da variação da ceratometria média e asfericidade e cálculo do erro do nomograma; (3) Aplicação dos algoritmos de aprendizagem de máquina e seleção de atributos; (4) Cálculo da variação da ceratometria média e da asfericidade prevista pelo algoritmo comparando com a variação do pré e pós operatório e cálculo do erro do algoritmo. Como resultado, o melhor valor do erro absoluto médio encontrado para asfericidade foi 0.19 e para ceratometria média foi 1.18. Comparando os valores do erro médio do nomograma e o erro médio do algoritmo, houve uma melhora de 0.11 para asfericidade e 0.09 para ceratometria média em relação ao nomograma atual, confirmando que a utilização de modelos computacionais é capaz de alcançar resultados mais precisos podendo contribuir para decisão cirúrgica na tentativa de melhorar a qualidade de visão de pacientes com ceratocone.

In this work is describe anatomy and physiology of the cornea. The following are the primary non-inflamatory degeneration of the cornea. Then describe the physical principles diagnostic devices for cornea – keratometer, pachymeter, Michelson interferometr and optical coherence tomography (OCT). At the end of the theoretical introduction is describes the principle of laser correction surgery – LASIK. The practical part is divided into two main objectives. The first task is propose an algorithm for automatic detection of corneal surface and then calculation of corneal thickness and size of the chamber angle in Matlab. The aim of the second task is image flap analysis for boundary detection.