This work aims to classify normal and carcinogenic cells in the oral cavity using two different approaches with an eye towards achieving high accuracy. The first approach extracts local binary patterns and metrics derived from a histogram from the dataset and is fed to several machine-learning models. The second approach uses a combination of neural networks as a backbone feature extractor and a random forest for classification. The results show that information can be learnt effectively from limited training images using these approaches. Some approaches use deep learning algorithms to generate a bounding box that can locate the suspected lesion. Other approaches use handcrafted textural feature extraction techniques and feed the resultant feature vectors to a classification model. The proposed method will extract the features pertaining to the images using pre-trained convolution neural networks (CNN) and train a classification model using the resulting feature vectors. By using the extracted features from a pre-trained CNN model to train a random forest, the problem of requiring a large amount of data to train deep learning models is bypassed. The study selected a dataset consisting of 1224 images, which were divided into two sets with varying resolutions.The performance of the model is calculated based on accuracy, specificity, sensitivity, and the area under curve (AUC). The proposed work is able to produce a highest test accuracy of 96.94% and an AUC of 0.976 using 696 images of 400× magnification and a highest test accuracy of 99.65% and an AUC of 0.9983 using only 528 images of 100× magnification images.

Skin lesions remain a significant global health issue, with their incidence rising steadily over the past few years. Early and accurate detection is crucial for effective treatment and improving patient outcomes. This work explores the integration of advanced Convolutional Neural Networks (CNNs) with Bidirectional Long Short Term Memory (BiLSTM) enhanced by spatial, channel, and temporal attention mechanisms to improve the classification of skin lesions. The hybrid model is trained to distinguish between various skin lesions with high precision. Among the models evaluated, the CNN (original architecture) with BiLSTM and attention mechanisms model achieved the highest performance, with an accuracy of 92.73%, precision of 92.84%, F1 score of 92.70%, recall of 92.73%, Jaccard Index (JAC) of 87.08%, Dice Coefficient (DIC) of 92.70%, and Matthews Correlation Coefficient (MCC) of 91.55%. The proposed model was compared to other configurations, including CNN with Gated Recurrent Units (GRU) and attention mechanisms, CNN with LSTM and attention mechanisms, CNN with BiGRU and attention mechanisms, CNN with BiLSTM, CNN with LSTM, CNN with BiGRU, CNN with GRU, standalone CNN, InceptionV3, Visual Geometry Group-16 (VGG16), and Xception, to highlight the efficacy of the proposed approach. This research aims to empower healthcare professionals by providing a robust diagnostic tool that enhances accuracy and supports proactive management strategies. The model’s ability to analyze high-resolution images and capture complex features of skin lesions promises significant advancements in early detection and personalized treatment. This work not only seeks to advance the technological capabilities in skin lesion diagnostics but also aims to mitigate the disease’s impact through timely interventions and improved healthcare outcomes, ultimately enhancing public health resilience on a global scale.

Detection and treatment of skin diseases is a complicated process as there are about 3000 skin diseases ^[39] . This adds complexity to the process of diagnosing skin diseases, highlighting the need for accurate detection to effectively treat the condition. Current deep learning-based skin detection tools typically concentrate on a limited subset of skin diseases from a vast spectrum, work on relatively small datasets, and rarely achieve a Top-5 accuracy above 70%. Ideally, disease detection systems should possess the capability to detect and classify skin diseases, taking into account various environmental and situational factors. To overcome these challenges in detecting skin diseases, a deep learning-based system is proposed, utilizing an ensemble method with existing architectures to enhance performance by making use of multiple models along with encoding to better adapt to varying inputs. The proposed deep learning-based system is trained on the DermNet dataset and uses a genetic algorithm optimized ensembling to enhance overall performance, resulting in a Top-5 accuracy of 74% on the DermNet dataset.

Renewable energy can lead to a sustainable future and solar energy is one the primary sources of renewable energy. Solar energy is harvested mainly by photovoltaic plants. Though there are a large number of solar panels, the economic efficiency of solar panels is not that high in comparison to energy production from coal or nuclear matter. The main risk involved in solar plants is the high maintenance cost involved in maintaining the plants. To help reduce this issue, automated solutions using Unmanned Aerial Vehicles (UAVs) and satellite imagery are proposed. In this research work, we propose a novel deep learning architecture for the segmentation of solar plant aerial images, which not only helps in automated solar plant maintenance, but can also be used for the area estimation and extraction of solar panels from an image. Along with this, we also propose a transfer learning-based model for the efficient classification of solar panel damage. Solar panel damage classification has a lot of applications. It can be integrated into monitoring systems, raising alerts when there is severe damage or damage of a certain type. The adaptive UNet model with Atrous Spatial Pyramid Pooling (ASPP) module that performed the dilated convolutions that we proposed achieved an overall accuracy of 98% with a Mean Intersection-Over-Union (IoU) Score of 95% and took under a second to process an image. Our classification model using Visual Geometry Group 19 (VGG19) as the backbone for feature extraction has achieved a classification accuracy of 98% with an F1 score of 99%, thus detecting the five classes of damage, including undamaged solar panels, in an efficient manner.

Ensuring sustainable and profitable agriculture is critical for addressing global food security challenges. This has resulted in the need for automation in plant health identification. However, this objective is hampered by the lack of efficient image-processing methods and the need for extensive datasets for training deep learning models for plant disease diagnosis. To overcome the need for extensive training data, the proposed Localized Normalized Difference Vegetation Index (LNDVI) uses zero-shot plant detection models such as Grounded Dino and state-of-the-art methods for image segmentation such as Segment Anything Model (SAM) are leveraged. This also expands the capabilities of the system to diagnose plant health beyond known plant species available as part of training set. The proposed system uses synthetic Normalized Difference Vegetation Index (NDVI) to estimate the chlorophyll content of the plant through RGB images alone instead of using the combination of RGB and near Infra-red (nIR) bands used in contemporary works. Since NDVI value is greatly affected by the amount of light present while the image is captured, we also present an irradiation estimation metric that uses CIE XYZ (Tristimulus values), Hue, Saturation and Value (HSV) and CIE LAB color spaces as well as correlated color temperatures, which automatically normalizes the NDVI threshold for health classification of the image, enabling a more precise analysis of plant health. Using the Grounding Dino provided an accuracy of 99.994% in terms of detecting plants from the phenotyping dataset. The segmentation of plant region in images is reported using Intersection over Union (IoU). While using the Segment Anything Model (SAM), an average accuracy of 95.884% was obtained for clustered plants while the average accuracy was even better at 97.031% for individual plants. Significant differences were observed for plant health classification while using Localized Normalized Difference Vegetation Index (LNDVI) approach when compared to NDVI.

The critical structure and nature of different bone marrow cells which form a base in the diagnosis of haematological ailments requires a high-grade classification which is a very prolonged approach and accounts for human error if performed manually, even by field experts. Therefore, the aim of this research is to automate the process to study and accurately classify the structure of bone marrow cells which will help in the diagnosis of haematological ailments at a much faster and better rate. Various machine learning algorithms and models, such as CNN + SVM, CNN + XGB Boost and Siamese network, were trained and tested across a dataset of 170,000 expert-annotated cell images from 945 patients’ bone marrow smears with haematological disorders. The metrics used for evaluation of this research are accuracy of model, precision and recall of all the different classes of cells. Based on these performance metrics the CNN + SVM, CNN + XGB, resulted in 32%, 28% accuracy, respectively, and therefore these models were discarded. Siamese neural resulted in 91% accuracy and 84% validation accuracy. Moreover, the weighted average recall values of the Siamese neural network were 92% for training and 91% for validation. Hence, the final results are based on Siamese neural network model as it was outperforming all the other algorithms used in this research.