image processing

An alternative to both the time-consuming traditional and the expensive three-dimensional (3D) methods for splint design is to use two-dimensional (2D) images. The present study utilized image processing to achieve an automatic and practical method of anthropometry measurement to design and build a personalized and remote cock-up splint. This method is applicable for patients unable to personally attend clinic appointments. The defined landmarks of the cock-up splint of 100 adult participants were measured manually. Each individual had a 2D image taken of their upper limb using a customized imaging device. The 2D image portions that corresponded to the manual measurements were then identified, and their sizes were retrieved in pixels using MATLAB software. To find equations between manual 3D measurements and 2D image processing ones, multiple linear regression analysis was performed on landmark variables. We were able to determine equations to estimate manual dimensions based on 2D image data. In the men’s group, we could predict the third finger length, forearm circumference at three levels, and the largest forearm circumference. In the women’s group, in addition to variables predicted for men, hand circumference at the distal palmar crease and first web levels, as well as arm circumference, could be predicted using the identified equations. Based on the findings, 2D image processing could be an appropriate method for designing personalized cock-up splints. Level of evidence: III

Wound detection is important in forensic science. The development of technology for automatic wound detection in the medical field has not been fully realized. This study aimed to compare the level of knowledge between medical students who utilized an application versus those who studied from textbooks.

An image processing-based application was developed using a Python program with the YOLO software, version 4 algorithm. The research utilized a cross-sectional study design. The participants were medical students undergoing clinical clerkships at the Forensic and Medicolegal Departments of the Faculty of Medicine, Public Health, and Nursing at Gadjah Mada University. Each group (textbook and application group) comprised at least 35 individuals. Both groups had identical study periods and answered the same questions. Participants then filled out a questionnaire about the application.

The mean scores of the post-test showed significant improvement in both groups compared to their pre-test scores (P=0.00). There was no significant difference in the post-test scores between the groups using the questionnaire and the application (P=0.207).

The application for wound type determination could serve as a preferable alternative to traditional textbooks for learning. It demonstrated an increase in learning outcomes that was significantly comparable to that achieved through textbook study.

Complexity metrics have been suggested to characterize treatment plans based on machine parameters such as multileaf collimator (MLC) position. Several complexity metrics have been proposed and related to the Intensity-modulated radiation therapy (IMRT) quality assurance results. This study aims to evaluate aperture-based complexity metrics on MLC openings used in clinicaland establish a correlation between plan complexity and the gamma passing rate (GPR) for the IMRT plans. We implemented the aperture-based complexity metric on MLC openings of the IMRT treatment plan for breast  and central nervous system (CNS) cases . The modulation complexity score (MCS), the edge area metric (EAM), the converted area metric (CAM), the circumference/area (CPA), and the ratio monitor unit MU/Gy are evaluated in this study. The complexity score was calculated using Matlab. The MatriXX Evolution was used for dose verification. The dose distribution was  analyzed using the OmniPro-I'mRT program  and the gamma index was assessed using two criteria: 3%/3 mm and 3%/2 mm. The correlation between the calculated complexity score and the GPR  is analyzed using SPSS. The complexity score calculated by MCS, EAM, CAM, CPA, and MU/Gy shows breast plan is more complex than the CNS plan. The results of the correlation test of the complexity metric and GPR show that only the EAM metric shows a good correlation with GPR for both cases. EAM strongly correlates with the gamma pass rate. The MCS, CAM, CPA, and MU/Gy have a weak correlation with the GPR.

Several approaches have been proposed to optimize the construction of an artificial intelligence-based model for assessing ploidy status. These encompass the investigation of algorithms, refining image segmentation techniques, and discerning essential patterns throughout embryonic development. The purpose of the current study was to evaluate the effectiveness of using U-NET architecture for embryo segmentation and time-lapse embryo image sequence extraction, three and ten hr before biopsy to improve model accuracy for prediction of embryonic ploidy status.

A total of 1.020 time-lapse videos of blastocysts with known ploidy status were used to construct a convolutional neural network (CNN)-based model for ploidy detection. Sequential images of each blastocyst were extracted from the time-lapse videos over a period of three and ten hr prior to the biopsy, generating 31.642 and 99.324 blastocyst images, respectively. U-NET architecture was applied for blastocyst image segmentation before its implementation in CNN-based model development.  

The accuracy of ploidy prediction model without applying the U-NET segmented sequential embryo images was 0.59 and 0.63 over a period of three and ten hr before biopsy, respectively. Improved model accuracy of 0.61 and 0.66 was achieved, respectively with the implementation of U-NET architecture for embryo segmentation on the current model. Extracting blastocyst images over a 10 hr period yields higher accuracy compared to a three-hr extraction period prior to biopsy.

Combined implementation of U-NET architecture for blastocyst image segmentation and the sequential compilation of ten hr of time-lapse blastocyst images could yield a CNN-based model with improved accuracy in predicting ploidy status.

The presence of pigmented skin lesions is a significant global concern in the prevention of skin cancer. Detecting skin cancer at an early stage is essential for proper management and effective treatment. This study aimed to combine image processing and data mining to develop an intelligent model to screen skin cancer from skin lesions.

The images were taken in a clinic by smartphone. Patients over 40 years of age participated in the study. During the segmentation phase, the lesions were separated from the original images through machine vision techniques. Various features such as symmetry, border irregularity, color variation, and diameter were extracted from the images, while some features were also obtained through face-to-face examination. Finally, a neural network was employed to classify whether the lesion was cancerous or non-cancerous. In addition, MATLAB version 2022 was considered to design the model.

The study results indicated excellent segmentation. Using a neural network-based model, skin lesions were classified with a high level of accuracy, with 98.4% accuracy and 97% sensitivity. The results indicated the designed model significantly screened skin cancer with high accuracy.

This model can help patients to manage self-care and become aware of their skin lesions before consulting a physician.

To investigate the success rate and quality of automatic airway segmentation using ultra-low dose CT (ULD-CT) images of different reconstruction algorithms.

Fifty two children who underwent chest ULD-CT were divided into three groups for analysis based on age: group A (n=13, age, 1-2years), group B (n=19, age, 3-6years) and group C (n=20, age, 7-13years). CT images were reconstructed with filtered back-projection (FBP), 50% adaptive statistical iterative reconstruction-Veo (50%ASIR-V), 100%ASIR-V, deep learning image reconstruction (DLIR) with low (DLIR-L), medium (DLIR-M), and high (DLIR-H) strengths. Subjective image quality was evaluated using a 5-point scale. CT value, noise, and sharpness of the trachea were measured. The VCAR software was used to automatically segment airways and reported the total volume. Segmentation success rates were recorded, and segmentation images were subjectively evaluated using a 6-point scale.

The average tracheal diameters were 8.53±1.88mm, 10.69±1.65mm, and 12.72±1.97mm, respectively for groups A, B, and C. The segmentation success rate depended on patient groups: group C reached 100%, while group A decreased significantly. In group A, 100%ASIR-V had the lowest rate at 7.69%, while DLIR-M and DLIR-H significantly improved the rate to 38.64% (P=0.03). For the segmented images, DLIR-H provided the lowest noise and highest subjective score while FBP images had the highest noise and 100%ASIR-V had the lowest overall score (P<0.05). There was no significant difference in the total airway volume among the six reconstructions.

The airway segmentation success rate in ULD-CT for children depends on the tracheal size. DLIR improves airway segmentation success rate and image quality.

The BEBIG Portio multi-channel applicator provides better target dose coverage and sparing organs-at-risk compared to a single-channel cylinder. However, artifacts and distortions of Portio in magnetic resonance images (MRI) have not yet been reported.

We aimed to quantify the artifacts and distortions in its 1.5-Tesla MR images before clinical use.

In this experimental study, we employed a gelatin-filled phantom to conduct our measurements. T2-weighted (T2W) images were examined for artifacts and distortions. Computed tomography (CT) images were used as a reference to assess image distortions. Artifact severity was measured by recording the full-width-at-half-maximum (FWHM) image pixel values at various positions along the length of the applicator/channels. CT and MRI-based applicator reconstruction accuracy were then compared, and signal-to-noise ratio (SNR) and contrast were also determined for the applicator images.

The applicator distortion level for the Portio applicator was less than the image spatial resolution (0.5±0.5 pixels). The average FWHM for the tandem applicator images was 5.23±0.39 mm, while it was 3.21±0.37 mm for all channels (compared to their actual diameters of 5.0 mm and 3.0 mm, respectively). The average applicator reconstruction difference between CT and MR images was 0.75±0.30 mm overall source dwell positions. The image SNR and contrast were both acceptable. 

These findings indicate that the Portio applicator has a satisfactory low level of artifacts and image distortions in 1.5-Tesla, T2W images. It may, therefore, be a promising option for MRI-guided multi-channel vaginal brachytherapy.

Malaria remains a significant global health problem, with a high incidence of cases and a substantial number of deaths yearly. Early identification and accurate diagnosis play a crucial role in effective malaria treatment. However, underdiagnosis presents a significant challenge in reducing mortality rates, and traditional laboratory diagnosis methods have limitations in terms of time consumption and error susceptibility. To overcome these challenges, researchers have increasingly utilized Machine Learning techniques, specifically neural networks, which provide faster, cost-effective, and highly accurate diagnostic capabilities.

This study aimed to compare the performance of a traditional neural network (NN) with a convolutional neural network (CNN) in the diagnosis and classification of different types of malaria using blood smear images. We curated a comprehensive malaria dataset comprising 1,920 images obtained from 84 patients suspected of having various malaria strains. The dataset consisted of 624 images of Falciparum, 548 images of Vivax, 588 images of Ovale, and 160 images from suspected healthy individuals, obtained from local hospitals in Iran. To ensure precise analysis, we developed a unique segmentation model that effectively eliminated therapeutically beneficial cells from the image context, enabling accurate analysis using artificial intelligence algorithms.

The evaluation of the traditional NN and the proposed 6-layer CNN model for image classification yielded average accuracies of 95.11% and 99.59%, respectively. These results demonstrate that the CNN, as a primary algorithm of deep neural networks (DNN), outperforms the traditional NN in analyzing different classes of malaria images. The CNN model demonstrated superior diagnostic performance, delivering enhanced accuracy and reliability in the classifying of malaria cases.

This research underscores the potential of ML technologies, specifically CNNs, in improving malaria diagnosis and classification. By leveraging advanced image analysis techniques, including the developed segmentation model, CNN showcased remarkable proficiency in accurately identifying and classifying various malaria parasites from blood smear images. The adoption of machine learning-based approaches holds promise for more effective management and treatment of malaria, addressing the challenges of underdiagnosis and improving patient outcomes.

X-rays are routinely utilized for different diagnostic purposes but there is always the risk of an inaccurate diagnosis. This systematic review was designed to investigate whether inverse grayscale mode increased diagnostic accuracy. From inception to February 2022, MEDLINE, Embase, Scopus, Web of Science, and CENTRAL were searched for studies comparing grayscale inversion diagnostic accuracy to the conventional method. Quality assessment was performed using the Quality Assessment of Diagnostic Accuracy Studies version 2 (QUADAS-2) tool. Eighteen studies were included with an overall patient population of 1704. The number of studies investigating each lesion are as follows, lung masses: 13, pneumothoraces: 4, bony lesions: 3, interstitial lung diseases: 3, orthopedic studies: 2, bullous lung disease: 1, pleural effusion: 1, urinary calculus: 1, and large vascular occlusion: 1. Two studies had an overall moderate risk of bias and the remainders had low risk. The combined mode, featuring the conventional mode with the addition of the inverse grayscale, demonstrated better performance or insignificant difference in comparison with the conventional mode in all studies except one, which showed lower sensitivity in detecting pulmonary nodules. Also, meta-analysis of 250 patients in four pulmonary nodule studies showed better area under the ROC curve (AUC) of inverse mode (0.83, 95% CI: 0.75,0.90) in comparison with conventional mode (0.80, 95% CI: 0.72,0.88). Application of inverse mode when using radiography for detection of pulmonary nodules might improve diagnostic accuracy. Also, the inverse/combined mode showed better performance for lesions other than pulmonary nodule in some studies. However, there was insufficient evidence to draw a consistent conclusion.

Automatic segmentation of the choroid on optical coherence tomography (OCT) images helps ophthalmologists in diagnosing eye pathologies. Compared to manual segmentations, it is faster and is not affected by human errors. The presence of the large speckle noise in the OCT images limits the automatic segmentation and interpretation of them. To solve this problem, a new curvelet transform‑based K‑SVD method is proposed in this study. Furthermore, the dataset was manually segmented by a retinal ophthalmologist to draw a comparison with the proposed automatic segmentation technique.

In this study, curvelet transform‑based K‑SVD dictionary learning and Lucy‑Richardson algorithm were used to remove the speckle noise from OCT images. The Outer/Inner Choroidal Boundaries (O/ICB) were determined utilizing graph theory. The area between ICB and outer choroidal boundary was considered as the choroidal region.

The proposed method was evaluated on our dataset and the average dice similarity coefficient (DSC) was calculated to be 92.14% ± 3.30% between automatic and manual segmented regions. Moreover, by applying the latest presented open‑source algorithm by Mazzaferri et al. on our dataset, the mean DSC was calculated to be 55.75% ± 14.54%.

A significant similarity was observed between automatic and manual segmentations. Automatic segmentation of the choroidal layer could be also utilized in large‑scale quantitative studies of the choroid.

The cataract is the most prevalent cause of blindness worldwide and is responsible for more than 51 % of blindness cases. As the treatment process is becoming smart and the burden of ophthalmologists is reducing, many existing systems have adopted machine-learning-based cataract classification methods with manual extraction of data features. However, the manual extraction of retinal features is generally time-consuming and exhausting and requires skilled ophthalmologists.

Convolutional neural network (CNN) is a highly common automatic feature extraction model which, compared to machine learning approaches, requires much larger datasets to avoid overfitting issues. This article designs a deep convolutional network for automatic cataract recognition in healthy eyes. The algorithm consists of four convolution layers and a fully connected layer for hierarchical feature learning and training.

The proposed approach was tested on collected images and indicated an 90.88 % accuracy on testing data. The keras model provides a function that evaluates the model, which is equal to the value of 84.14 %, the model can be further developed and improved to be applied for the automatic recognition and treatment of ocular diseases.

This study presented a deep learning algorithm for the automatic recognition of healthy eyes from cataractous ones. The results suggest that the proposed scheme outperforms other conventional methods and can be regarded as a reference for other retinal disorders.

The primary aim of this study was to assess the reliability of the ten -segment classification system proposed (TSC) by Krause et al. and see how it compares with the traditionally used Schatzker classification, AO classification system, and Luo’s “Three columns” classification (ThCC) system. The second aim of this study was to assess the inter-observer reliability of the above classifications based on professional experience by comparing the entry level of residents (1 year into postgraduation), senior residents (1 year after postgraduation completion), and faculty (>10 years after postgraduation completion). 50 TPFs were classified by a 10-segment classification system, and its intra-observer (at 1-month interval) and inter-observer reproducibility was checked using k values by three different groups with varying levels of experience (Group I, II, and III comprised of 2 juniors residents, senior residents and consultants each), and the same was compared for three other common classification systems (Schatzker, AO and 3 –column). 10-segment classification showed least k for both inter-observer (0.08) and intra-observer (0.03) reliability. Highest individual inter-observer (k= 0.52) and intra-observer reliability (k= 0.31) was for Schatzker classification in Group I. Lowest individual inter-observer and intra-observer reliability was seen for 10-segment classification (k= 0.07) and AO classification system (k= -0.03) respectively. 10-segment classification showed the lowest k for both inter-observer and intra-observer reliability. The inter-observer reliability for the Schatzker, AO, and 3- column classifications reduced with increasing experience of the observer (JR>SR>Consultant). A possible reason could be a more critical evaluation of the fractures with increasing seniority. Level of evidence: I

The TBS-derived image processing method, based on the observer's diagnosis, has been developed in the current investigation. Image parametrization is proposed for both novel description and convergent shreds of evidence.

Condensed X-ray images of the esophageal timed barium swallow (TBS) provide substantial implications for elucidating the pathophysiological dimensions of esophageal motility disorders.

Throughthe simultaneous study on TBS and high-resolution manometry (HRM) findings, we performed a retrospective cohort study on 252 patientsfrom March 2018 to October 2019. Interventions, irrelevant information, and insufficient patient data were excluded. Only subjects with adequate data and acceptable test accuracy were considered for participation. We reviewed 117 Dicom (digital imaging and communications in medicine) X-ray images from patients with confirmed diagnoses of achalasia type II, esophagogastric junction outflow obstruction (EGJOO), or non-achalasia.

The results suggested a cut-off level of 47% in DDi (dilated diameter index) as a sign of the dilated body. In achalasia type II patients (n=66 images), the mean DDi was 55.6%. Our method presented a sensitivity of 95% and a specificity of 93% compared to images of the non-achalasia findings. The mean DDi in EGJOO patients was 50.4%, according to the 27 images. Moreover, results from EGJOO patients provided a sensitivity of 85% and specificity of 87%.

TBS is an efficacious method and a prominent component in the process of achalasia diagnosis. Standard parametrization might develop radiological exports proposed by DDi. Our method could assist in obtaining a non-invasive medical diagnosis and help advance diagnostic reports to identify achalasia subtypes somewhat earlier. To the best of our knowledge, this interface is an innovative parametrization for TBS image review.

Accurate and early diagnosis of internal root resorption is essential for determining the outcome of treatment and prognosis. Several digital processing algorithms have been introduced for the diagnosis of internal root resorption. The aim of the present study was to evaluate the effect of gamma values in the detection of internal root resorption in cone-beam computed tomography (CBCT) images.

A total of 45 healthy extracted single-rooted teeth were selected for the study. The teeth were mesiodistally sectioned at the central groove of their occlusal surface using a diamond disc (0.1 mm diameter). Internal root resorption was simulated at the cervical, middle, and apical regions of a root canal. CBCT images were prepared by gamma values (low, medium, and high modes). Data were analyzed using McNemar tests, Cohen’s kappa coefficient, and the receiver operating characteristic curve. A P value less than 0.05 (typically≤0.05) was considered statistically significant.

Sensitivity and overall accuracy of CBCT images with all gamma modes (low, medium, and high) were high, and they were slightly higher in the high gamma mode (Sensitivity: 94% and overall accuracy: 100%) than in two gamma modes (low and medium). The sensitivity and specificity of high and medium gamma modes in three regions of the root canal (cervical, middle, and apical) were at optimal thresholds compared to the low gamma mode.

Gamma modes, particularly the high gamma mode, in CBCT imaging can be adopted as a promising processing filter for the detection of internal root resorption.