Development of a deep learning based image processing tool for enhanced organoid analysis Scientific Reports

The color variations of histopathology images (i.e., domain shift) notably affect the amplitude spectrum of images, while phase-related components carry more informative content. In light of this, we hypothesize that incorporating frequency information has the potential to enhance both the discriminability and knowledge transferability of the domain adversarial networks. The FFT-Enhancer module, with its straightforward calculation and minimal computational burden, emerges as an ideal candidate for this purpose. To better understand the progress of the algorithm, additional models were trained for comparison, including DeeplabV3, DeeplabV3 + , FPN, Linknet, PAN, Pspnet, and UNet +  + alongside Transformer + UNet. The training strategies for all models were kept consistent to ensure a fair comparison. Specifically, all models were trained on the same dataset, using identical batch sizes, learning rates, and optimization methods.

Experimental results show that the Transformer + UNet hybrid model achieves an accuracy of 95.57% in lithology identification tasks, while the optimized ResNet model achieves an accuracy of 96.13% in rock weathering degree determination. Additionally, the average relative error in tunnel face strength detection results is only 9.33%, validating the feasibility and effectiveness of this method in practical engineering applications. The multi-model neural network system developed in this study significantly enhances prediction accuracy and efficiency, providing robust scientific decision support for tunnel construction, thereby improving construction safety and economy. V.G.G. is a convolutional neural network (CNN) architecture which addresses the important aspect of depth in deep networks and uses small convolutional filters, allowing the model to have a large number of weight layers.

ORIGINAL RESEARCH article

The study proposes an innovative AI-based solution to accurately identify handloom “gamucha”s amidst powerloom counterparts, aiming to combat misrepresentation in the market. By meticulously curating a benchmarked image dataset and developing a custom deep learning model, the study pioneers an automated recognition system capable of differentiating between handloom and powerloom “gamucha”s with high accuracy. The proposed model not only streamlines the identification process but also offers a user-friendly mobile application for seamless deployment, facilitating wider adoption and practical implementation.

After passing through the SENet (right-side colored C), different colors represent varying weights, altering the importance of each feature channel. The core idea is to classify feature weights based on loss through a fully connected network. Although it slightly increases parameters and computational complexity, it achieves better learning results21,22. In the biomedical field, image processing based on artificial intelligence (AI) can enable effective image data analysis6.

Top 9 Machine Learning Challenges in 2024

As we delve into the creative and security spheres, Prisma and Sighthound Video showcase the diverse applications of image recognition technology. Microsoft Seeing AI and Lookout by Google exemplify the profound ChatGPT impact on accessibility, narrating the world and providing real-time audio cues for individuals with visual impairments. Runway ML emerges as a trailblazer, democratizing machine learning for creative endeavors.

• BURANO Version 2.0 will also add 1.8x de-squeeze setting as well as additional high frame rate (S & Q) modes including 66, 72, 75, 88, 90, 96, 110 fps.
• In today’s teaching practice, there are often phenomena, such as reading from books and leaving textbooks, which seriously affect the improvement of teaching quality.
• In the slightly weathered stage, the rock structure begins to deteriorate, with noticeable changes in color and mineral composition.
• It gauges the differences and commonalities between texts with basic calculation methods, including string matching and word matching.

Advertise with TechnologyAdvice on Datamation and our other data and technology-focused platforms. Finally, Version 2.0 will add breathing compensation and image stabilization metadata in X-OCN. And while the Sony BURANO might ChatGPT App have been the most talked about new Sony cinema camera this year, Sony has built-out an impressive stable of high-quality cinema cameras which now includes its VENICE 2, FX6, FX3, FX30 and BURANO cameras these days.

A geometric approach for accelerating neural networks designed for classification problems

Moreover, it recognizes that patches within an image have varying degrees of importance to its subtype, and their contributions are calculated using an attention mechanism. Visualizing the feature spaces of Base, CNorm, and AIDA for datasets related to (a) Ovarian Cancer, b Pleural Cancer, and (c) Bladder Cancer. The results, shown in Table 4, indicate that the average error values for Tunnel 1, Tunnel 2, and Tunnel 3 are 9.772%, 8.844%, and 9.58%, respectively. These results demonstrate a high consistency between laboratory testing and on-site identification methods in determining the strength of the tunnel face surrounding rock, validating the effectiveness of the correction factor method. This method allows the construction team to more accurately assess the strength of the surrounding rock, thereby optimizing construction plans and improving construction safety and efficiency.

Datamation’s focus is on providing insight into the latest trends and innovation in AI, data security, big data, and more, along with in-depth product recommendations and comparisons. Learn the latest news and best practices about data science, big data analytics, artificial intelligence, data security, and more. In e-commerce product recommendation systems, AI categorizes products based on user behavior, preferences, and purchase history. It makes use of collaborative filtering or content-based filtering techniques to match users with relevant products. Version 2.0 also offers monitoring improvements, including standardized SDI video output for monitoring across X-OCN and XAVC and an improved on-screen display that places camera status information outside of the image.

For CXP, decreasing the window width and field of view increases the model’s average score corresponding to white patients and decreases the average score corresponding to Asian and Black patients. The factors are combinatorial, with changes of −68.3 ± 0.6%, −58.0 ± 1.0%, +33.0 ± 0.5% in the average Asian, Black, and white prediction scores, respectively, when changing each parameter by 20%. These patterns are similar in the MXR dataset for Asian and white prediction scores (Fig. 2b).

● Firstly, this review organized the standard data sets and evaluation indicators. You can foun additiona information about ai customer service and artificial intelligence and NLP. The list of datasets and their evaluation methods are in-depth and highlighted from different literature from recent years. Building upon the preceding analysis, this work aims to employ the TSM method to assess the similarity between the classroom discourse of online courses in secondary schools and the teaching courseware. Subsequently, it analyzes the disparities and similarities between the two, evaluates the degree of “Scripted Teaching” by teachers, and offers insights for enhancing classroom discourse. As shown in Table 3, the length of sentences is a fundamental indicator for evaluating the complexity of classroom discourse.

This approach can greatly reduce training time, improve the accuracy and reliability of the model, and utilize the computing resources of multiple computing nodes to process large-scale datasets. Model parallelism is where ai based image recognition the distributed nodes store different parameters with the same input layer. Data parallelism is the exact opposite, where each distributed node stores the same parameters with different data in the input layer.

Table 3 Advantages, disadvantages, and applicable situations of single-stage Object detection algorithms. In one-stage detectors, the class imbalance of foreground and background is the main reason for the convergence of network training. During the training phase, Focal Loss avoids many simple negative examples and focuses on hard training samples.

Large libraries of pretrained CNNs that can be used to ‘donate’ their front ends are available as open source from the major cloud providers hoping to capture the cloud compute revenue, and also from academic sources. Power companies rely on human labor to visually inspect their towers covering large geographic areas. These manual inspections are notorious for being expensive, risky and slow, especially when the towers are spread over mountainous or inaccessible terrain. As a solution, power companies can transform inspection jobs by deploying drones with cameras and adopting PowerAI Vision.

Central to adversarial domain adaptation is the feature extractor network, which operates with the dual objective of learning representations that are both discriminative for the primary task and invariant to domain shifts. Several studies have investigated the effectiveness of adversarial networks for domain adaptation in the realm of histopathology images13,35,36. In the works by Lafarge et al.35 and Otalora et al.36, adversarial networks are employed for domain adaptation tasks concerning mitosis detection in breast cancer histopathology images and Gleason pattern classification in prostate cancer. These approaches involve incorporating the adversarial network with color augmentation and stain normalization. However, it’s important to note that this integration may not be optimal, as the adversarial network should ideally demonstrate robust performance without relying on color normalization support.

When conducting machine learning training tasks, we need to preprocess the image data in the database to enhance data quality and consistency, thereby improving the accuracy and reliability of the analysis results. The first step in preprocessing is to crop the images to remove irrelevant details, ensuring the focus remains on the rock samples, which maintains dataset uniformity. In tunnel construction, obtaining clear photos of the tunnel face is crucial for image training.

Some of the important handcrafted features for identification “gamucha” handloom. Addressing this issue is crucial to preserving the handloom heritage and supporting the livelihoods of the weaver community4,5,6. The Declaration of Helsinki and the International Ethical Guidelines for Biomedical Research Involving Human Subjects were strictly adhered to throughout this study. All protocols for this study, including the waiver of consent, have been approved by the University of British Columbia/BC Cancer Research Ethics Board. The refined classification scheme that leverages AI screening as a supplementary stratification mechanism within the NSMP molecular subtype. In the frequency domain, the predominant effect of domain shift is amplitude changes, while the phase information is nearly identical (Supplementary Fig. 3).

Before the convolution operation, the feature image is normalized and preprocessed, and the activation function is used for nonlinear transformation. Similarly, the conversion layer also incorporates \(1 \times 1\) convolution, reducing the number of model feature maps determined by the compression coefficient. To solve these problems, on the one hand, the study proposes an improved strategy for feature reuse in dense convolutional network (DenseNet), which reduces the complexity of the model.

DL CNN outperformed all the ML classifiers (SVM, KNN, DT, and RF) and achieved accuracy rates of 99.58% for rice and 97.66% for potato leaf diseases in this research. Segmentation is a fundamental technique used in agricultural science, wherein an image is meticulously divided into its components. The primary goal is to analyze each object in more detail, extracting beneficial features that might enhance our understanding and knowledge (Jafar et al., 2022). Clustering, anomaly detection, and association rule mining are examples of unsupervised learning algorithms that extract meaningful insights and relationships from data. These algorithms are employed to segment markets, provide personalized product recommendations, detect outliers in data, and identify communities in social networks. Model deployment entails integrating the models into operational environments or business workflows, allowing for the real-world application of the classification outcomes.

The best accuracy result was obtained with VGG16 with 98%, while the F-score value of the same transfer learning model was 97%, the Area Under the Curve (AUC) value was 99%, the recall value was 98%, and the precision value was 98%. CNN architecture and CNN-based transfer learning models are very important for human health in early diagnosis and rapid treatment of such diseases. Adversarial networks15 have become prevalent in addressing the challenge of domain adaptation. By leveraging adversarial training techniques, domain adaptation models aim to bolster the model’s adaptability to diverse data distributions encountered during deployment. This is achieved through the alignment of feature distributions between the source and target domains.