Demystifying evidential Dempster Shafer-based CNN architecture for fetal plane detection from 2D ultrasound images leveraging fuzzy-contrast enhancement and explainable AI

Ultrasound imaging is a valuable tool for assessing the development of the fetal during pregnancy. However, interpreting ultrasound images manually can be time-consuming and subject to variability. Automated image categorization using machine learning algorithms can streamline the interpretation process by identifying stages of fetal development present in ultrasound images. In particular, deep learning architectures have shown promise in medical image analysis, enabling accurate automated diagnosis. The objective of this research is to identify fetal planes from ultrasound images with higher precision. To achieve this, we trained several convolutional neural network (CNN) architectures on a dataset of 12400 images. Our study focuses on the impact of enhanced image quality by adopting Histogram Equalization and Fuzzy Logic-based contrast enhancement on fetal plane detection using the Evidential Dempster-Shafer Based CNN Architecture, PReLU-Net, SqueezeNET, and Swin Transformer. The results of each classifier were noteworthy, with PreLUNet achieving an accuracy of 91.03%, SqueezeNET reaching 91.03% accuracy, Swin Transformer reaching an accuracy of 88.90%, and the Evidential classifier achieving an accuracy of 83.54%. We evaluated the results in terms of both training and testing accuracies. Additionally, we used LIME and GradCam to examine the decision-making process of the classifiers, providing explainability for their outputs. Our findings demonstrate the potential for automated image categorization in large-scale retrospective assessments of fetal development using ultrasound imaging.

Aggregating Different Scales of Attention on Feature Variants for Tomato Leaf Disease Diagnosis from Image Data: A Transformer Driven Study

Tomato leaf diseases can incur significant financial damage by having adverse impacts on crops and, consequently, they are a major concern for tomato growers all over the world. The diseases may come in a variety of forms, caused by environmental stress and various pathogens. An automated approach to detect leaf disease from images would assist farmers to take effective control measures quickly and affordably. Therefore, the proposed study aims to analyze the effects of transformer-based approaches that aggregate different scales of attention on variants of features for the classification of tomato leaf diseases from image data. Four state-of-the-art transformer-based models, namely, External Attention Transformer (EANet), Multi-Axis Vision Transformer (MaxViT), Compact Convolutional Transformers (CCT), and Pyramid Vision Transformer (PVT), are trained and tested on a multiclass tomato disease dataset. The result analysis showcases that MaxViT comfortably outperforms the other three transformer models with 97% overall accuracy, as opposed to the 89% accuracy achieved by EANet, 91% by CCT, and 93% by PVT. MaxViT also achieves a smoother learning curve compared to the other transformers. Afterwards, we further verified the legitimacy of the results on another relatively smaller dataset. Overall, the exhaustive empirical analysis presented in the paper proves that the MaxViT architecture is the most effective transformer model to classify tomato leaf disease, providing the availability of powerful hardware to incorporate the model.

Role of Transabdominal Ultrasonogram for Evaluation of Placental Maturity in Relation with Fetal Gestational Age

In this study our main goal is to evaluate the role of ultrasonography (USG) for determination of placental maturity and fetal gestational age. This cross-sectional study was done at the Department of Radiology and Imaging (USG section) Mymensingh Medical College and Hospital, Mymensingh from July 2008 to June 2010, where 60 patients included in this study, who was attending in the department of Radiology & Imaging for transabdominal ultrasonography with early and late pregnancy related complaints. In this study among the 60 patients, the youngest one was eighteen years and oldest one was thirty-nine years old age. Thirty five percent (35.0%) patients were from 26-30 years age group. About 33 patients out of 60(55.0%) were house wives. Most of the patients were presented with amenorrhea (65.0%) 39 out of 60. Among 60 patients, 20 patients (33.3%) were in gestational age within 12-28 weeks, 20 patients (33.3%) were in gestational age within 29-36 weeks and 20 patients (33.3%) were within >36 weeks gestational age. Among them, 20 patients (33.3%) had grade III placenta, 20 patients (33.3%) had grade II placenta, 12 patients (20%) had grade I placenta and 08 patients (13.3%) had grade 0 placenta. Out of 60 patients, 18 patients (30.0%) were in high risk group and 70.0% were normal. Six (6) patients (10.0%) suffered from HTN, 3 patients had RH negative (5.0%) blood group, 3(5.0%) patients suffered from APH, 3 patients suffer from DM and 3 from IUGR. In this study showed hypertension and IUGR had strong correlation with accelerated placental maturation. Maternal DM and Rh sensitization were associated with delayed maturation of the placenta. This study concludes that, USG appears to be the best imaging modality for the evaluation of placenta and its grading. USG is relatively less expensive and it is good considering the diagnostic accuracy in pregnancy profile. It is noninvasive procedure without any radiation hazards and better visualization of the lesion in different section, but this study is not a complete reflection of overall incidence and statistics regarding the ailment in our country. For this a more extensive study over a longer period covering different section of society is very much needed for better outcome.

Doppler Assessment of Hepatic Venous Waves for Evaluation of Large Varices in Cirrhotic Patient

This cross sectional study was conducted to evaluate the role of doppler ultrasonography of hepatic venous waves for evaluation of large varices in cirrhotic patients from July 2013 to June 2015 in Mymensingh Medical College Hospital, Mymensingh, Bangladesh. Patients getting admitted in the ward with a diagnosis of cirrhosis were enrolled in the study and purposive sampling technique was adopted. The sample size was 43. Data were collected by face to face interview and some data were gathered by records review and analyzed with the help of SPSS windows version - 12 software programs. Statistical significance was set at P<0.05 and confidence interval set at 95%. The research protocol was approved by the local ethical committee. Esophagogastroduodenoscopy is the gold standard for the diagnosis of esophageal varices. If the gold standard is not available, other possible diagnostic steps would be Doppler ultrasonography of the blood circulation (not endoscopic ultrasonography). Although and it can certainly demonstrate the presence of varices. In 60.47% of patient's monophasic wave pattern was seen and in 39.53% of cases biphasic & triphasic wave pattern were detected. Endoscopic examination was performed in all selected patients. In this study, 67.44% is large varices, 32.66% is small varices. Chi Square test was done for hypothesis testing and it was found significant (<0.05) and it indicates monophasic wave in Doppler USG signifies large varices. This test was also done to find out whether any significant difference of hepatic venous waveform in between male and female but it was not significant (>0.05). Diagnostic performance of USG for evaluation of varices showed, Sensitivity: 86.2%, Specificity: 92.85%, PPV: 96%, NPV: 76.47%, Accuracy 88%. Correlation co-efficient was 0.0064 which indicates moderately positive correlation in between monophasic hepatic venous waveform pattern by Doppler USG and large varices in oesophagogastroduodenoscopic findings Normal hepatic wave form shows triphasic pattern. Loss of this pattern in cirrhosis is mainly due to loss of compliance of liver. In conclusion, the loss of triphasic pattern of hepatic wave form is highly sensitive in predicting the presence of large varices in cirrhotic patients and this doppler parameter may be used as a non-invasive test for cirrhotic patients, who wish to avoid upper GI endoscopy. Further studies using a combination of various doppler parameters are needed to create indices with a better predictive value.

Correlation between Umbilical Cord Diameter and Cross Sectional Area with Gestational Age and Foetal Anthropometric Parameters

The objective of the study was to find out correlation between umbilical cord diameter, cross sectional area with gestational age and foetal anthropometric parameters. This cross sectional study was conducted among healthy women between the 24(th) and 40(th) completed weeks of a normal pregnancy in the Department of Radiology & Imaging, Mymensingh Medical College Hospital, Mymensingh during the study period, from July 2009 to June 2011. A total of 230 consecutive normal pregnancy patients were included in the study. The diameter & cross-sectional area of the umbilical cord were measured on a plane adjacent to the junction of the umbilical cord and the fetal abdomen, in cross-section, with maximum magnification of the image. The cord was manually circled, and it's cross sectional areas was automatically calculated by the ultrasonograph. The mean±SD age was 24.3±4.7 years with range from 19 to 36 years. The mean gestational age was 32.1±4.5 weeks and more than a half (56.4%) of the pregnant women were nulliparas. A positive significant (p<0.001) correlation were found between umbilical cord diameter with bi-parietal diameter (r=0.548); head circumference (r=0.411); abdominal circumference (r=0.444); femur length (r=0.366) and gestational age gestation age (r=0.643). Similarly, a significant (p<0.001) positive week correlation were found between umbilical cross sectional area with bi-parietal diameter (r=0.3303); head circumference (r=0.3202); abdominal circumference (r=0.2651); femur length (r=0.3307) and gestation age (r=0.4051). A positive significant better correlation was found with umbilical cord diameter than cross sectional area with foetal anthropometric parameters.