Rice is the primary staple food in many Asian countries, and ensuring the quality of rice crops is vital for food security. Effective crop management depends on the early and precise detection of common rice diseases such as bacterial blight, blast, brown spot, and tungro. This work presents a convolutional neural network model for classifying rice leaf disease. Four distinct diseases, bacterial blight, blast, brown spot, and tungro, are the main targets of the model. Previously, leaf pathologies in crops were mostly identified manually using specialized equipment, which was time-consuming and inefficient. This study offers a remedy for accurately diagnosing and classifying rice leaf diseases through deep learning techniques. Using this dataset, the proposed CNN model was trained to identify complex patterns and attributes linked to each disease using its deep learning capabilities. This CNN model achieved an exceptional accuracy of 99.99%, surpassing the benchmarks set by existing state-of-the-art models. The proposed model can be a useful diagnostic and early warning system for rice leaf diseases. It could help farmers and other agricultural professionals reduce crop losses and enhance the quality of their yields.

The perception of animal emotions is key to enhancing veterinary practice, human–animal interactions, and protecting domesticated species’ welfare. This study presents a unique emotion classification deep learning-based approach for pet animals. The actual and emotional status of dogs and cats have been classified using a modified EfficientNetB5 model. Utilizing a dataset of images classified into four different emotion categories—angry, sad, happy, and neutral—the model incorporates sophisticated feature extraction methods, such as Dense Residual Blocks and Squeeze-and-Excitation (SE) blocks, to improve the focus on important emotional indicators. The basis of the second strategy is EfficientNetB5, which is known for providing an optimal balance in terms of accuracy and processing capabilities. The model exhibited robust generalization abilities for the subtle identification of emotional states, achieving 98.2% accuracy in training and 91.24% during validation on a separate dataset. These encouraging outcomes support the model’s promise for real-time emotion detection applications and demonstrate its adaptability for wider application in ongoing pet monitoring systems. The dataset will be enlarged, model performance will be enhanced for more species, and real-time capabilities will be developed for real-world implementation.

Background: Nuclei segmentation is the first stage of automated microscopic image analysis. The cell nucleus is a crucial aspect in segmenting to gain more insight into cell characteristics and functions that enable computer-aided pathology for early disease detection, such as prostate cancer, breast cancer, brain tumors, and other diagnoses. Nucleus segmentation remains a challenging task despite significant advancements in automated methods. Traditional techniques, such as Otsu thresholding and watershed approaches, are ineffective in challenging scenarios. However, deep learning-based methods exhibit remarkable results across various biological imaging modalities, including computational pathology. Methods: This work explores machine learning approaches for nuclei segmentation by evaluating the quality of nuclei image segmentation. We employed several methods, including K-means clustering, Random Forest (RF), Support Vector Machine (SVM) with handcrafted features, and Logistic Regression (LR) using features derived from Convolutional Neural Networks (CNNs). Handcrafted features extract attributes like the shape, texture, and intensity of nuclei and are meticulously developed based on specialized knowledge. Conversely, CNN-based features are automatically acquired representations that identify complex patterns in nuclei images. To assess how effectively these techniques segment cell nuclei, their performance is evaluated. Results: Experimental results show that Logistic Regression based on CNN-derived features outperforms the other techniques, achieving an accuracy of 96.90%, a Dice coefficient of 74.24, and a Jaccard coefficient of 55.61. In contrast, the Random Forest, Support Vector Machine, and K-means algorithms yielded lower segmentation performance metrics. Conclusions: The conclusions suggest that leveraging CNN-based features in conjunction with Logistic Regression significantly enhances the accuracy of cell nuclei segmentation in pathological images. This approach holds promise for refining computer-aided pathology workflows, potentially leading to more reliable and earlier disease diagnoses.

Accurate detection of fractures in X-ray images is important to initiate appropriate medical treatment in time—in this study, an advanced combined attention CNN model with multiple attention mechanisms was developed to improve fracture detection by deeply representing features. Specifically, our model incorporates squeeze blocks and convolutional block attention module (CBAM) blocks to improve the model’s ability to focus on relevant features in X-ray images. Using computed tomography X-ray images, this study assesses the diagnostic efficacy of the artificial intelligence (AI) model before and after optimization and compares its performance in detecting fractures or not. The training and evaluation dataset consists of fractured and non-fractured X-rays from various anatomical locations, including the hips, knees, lumbar region, lower limb, and upper limb. This gives an extremely high training accuracy of 99.98 and a validation accuracy 96.72. The attention-based CNN thus showcases its role in medical image analysis. This aspect further complements a point we highlighted through our research to establish the role of attention in CNN architecture-based models to achieve the desired score for fracture in a medical image, allowing the model to generalize. This study represents the first steps to improve fracture detection automatically. It also brings solid support to doctors addressing the continued time to examination, which also increases accuracy in diagnosing fractures, improving patients’ outcomes significantly.

Background/Objectives: Spatial transcriptomics (ST) is costly and rarely available in routine care, whereas hematoxylin and eosin (H&E) slides exist for almost every tumour. Many published estimates of expression predicted from H&E use evaluations in which the test patient's tissue was also seen in training. We asked how well expression can be predicted for unseen patients. Methods: We built REGEN, an ensemble of a frozen Phikon encoder and fine-tuned DenseNet121 and ResNet50 networks, with k-nearest-neighbour retrieval on the two CNN branches. We evaluated it by leave-one-patient-out validation on two public cohorts, Her2ST (HER2-positive breast cancer, 8 patients) and cSCC (cutaneous squamous cell carcinoma, 4 patients), against a corrected reimplementation of a joint-contrastive model (MViTGene). Results: On the 50 most highly expressed genes (HEG), REGEN reached a mean Pearson correlation of 0.158 on Her2ST and 0.210 on cSCC, against 0.035 and 0.030 for the baseline. Over all 1,000 panel genes, correlations were much lower and concentrated in strongly expressed immune, stromal and epidermal genes. Choosing ensemble weights without the evaluated patient changed HEG by at most 0.005, whereas a gene panel selected on all sections differed in up to 37% of its genes without the held-out patient. On cSCC the correlation fell from 0.21 within patients to 0.05 after pooling patients. REGEN did not reach published mclSTExp values, whose protocol we could not verify. Conclusions: Predicted expression for unseen patients is modest and uneven across genes and patients, and does not support clinical use. With 8 and 4 patients the findings are exploratory; the contribution is a protocol-controlled comparison, not a new architecture.

Brain tumour classifiers are often reported as near-perfect and explainable without checking patient structure or annotations. We reproduced a recent attention framework (SSPANet) on the Figshare dataset (3064 T1-weighted contrast-enhanced slices, 233 patients). Its image-level split is fully patient-leaky; every configuration exceeds 98% accuracy, which drops by 5–9 points on the patient-wise split. Under a fixed training recipe, seven attention modules stayed within ±0.63 points of the plain backbone, whereas the recipe changed accuracy by up to 4.7 points. As a case study, TG-SSPANet uses expert masks to supervise attention, train a segmentation decoder and gate evidence pooling. With ResNet50, it reached 96.8% slice-level and, averaging slice probabilities, 98.3% patient-level accuracy (84.5% if every slice had to be correct; 15% of patients had discordant slices), within the confidence intervals of simpler mask-using pipelines; its attention module added no measurable benefit. Its Grad-CAM++ maps hit the tumour in 81% of held-out slices, but this follows from its decoder by construction, not the classifier’s reasoning. After tumour inpainting, whole-slice models kept 80–86% accuracy, and shuffling the remaining tissue left 60–74%, so location, perilesional cues and tissue statistics cannot be separated. Patient-wise evaluation, same-recipe controls and mask-based faithfulness tests should precede explainability claims.