DenseNet
PulseAugur coverage of DenseNet — every cluster mentioning DenseNet across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
New research reveals flaws in AI benchmarking for medical image and EHR data
A new paper published on arXiv highlights significant issues with the evaluation of class-imbalance strategies in multimodal medical classification tasks. The research found that benchmarks using record-level splits on …
-
AI models evaluated for IoT DDoS detection under resource constraints
A new research paper evaluates several pre-trained convolutional neural network architectures for detecting distributed denial-of-service (DDoS) attacks in resource-constrained Internet of Things (IoT) environments. The…
-
New GRASP method enhances text-to-image transformers for rare conditions
Researchers have developed GRASP (Guided Residual Adapters with Sample-wise Partitioning), a novel method to improve the performance of text-to-image transformers in long-tail settings. This approach addresses gradient …
-
Chest X-ray AI models show limited real-world performance in multinational study
A new retrospective study evaluated ten Chest X-ray AI models across twelve datasets from six countries, including both public benchmarks and private institutional cohorts. The research found that while models like MAVL…
-
Deep learning methods developed for subcortical brain structure segmentation
This research paper introduces two deep learning algorithms, DenseMedic and Alternate Connected Neural Network (ACNN), designed for the semantic segmentation of subcortical brain structures in MR images. DenseMedic util…
-
New AI method improves weather recognition for autonomous driving
Researchers have developed a new method called Weather-Aware Adversarial Discriminative Domain Adaptation (WA-ADDA) to improve weather recognition from street-view images, a critical task for autonomous driving systems.…
-
New AI calibration method tackles shortcut learning in classifiers
Researchers have proposed a new approach to mitigate shortcut learning in AI classifiers by reframing the problem as one of calibration. Their methods, an in-processing regularizer and a post-hoc recalibration step, aim…
-
AI models achieve high accuracy in retinal disease classification and vessel segmentation
Researchers have developed a novel two-pipeline framework for analyzing retinal fundus images, combining disease classification with blood vessel segmentation. The framework fine-tuned eight ImageNet-pretrained CNNs for…
-
Deep learning models benchmarked for lung cancer histopathology analysis
Researchers have developed a two-stage deep learning framework for analyzing lung cancer histopathology images. The framework systematically compares state-of-the-art architectures for both tissue classification and reg…
-
Soft-Attention mechanism boosts skin cancer classification in deep neural networks
Researchers have demonstrated that incorporating a Soft-Attention mechanism into deep neural network architectures can significantly improve their performance in classifying skin lesions. By enabling networks to focus o…
-
VGG16 leads Alzheimer's detection in MRI scans across ten CNNs
Researchers have benchmarked ten different convolutional neural network (CNN) architectures for detecting Alzheimer's disease from single-view MRI scans. The study utilized a transfer learning and fine-tuning pipeline o…
-
New deep learning model Dense-Cast forecasts precipitation with high accuracy
Researchers have developed Dense-Cast, a new lightweight deep learning model designed for short-term precipitation nowcasting. The model integrates DenseNet architecture, residual connections, and transformer encoders t…
-
Chest X-ray ML performance heavily influenced by evaluation references, study finds
A new research paper published on arXiv explores the critical impact of evaluation references on the performance metrics of machine learning models used for chest X-ray analysis. The study highlights that commonly used …
-
New dataset and deep learning model estimate human weight and height from images
Researchers have developed a method for estimating human weight and height from single images captured in everyday settings. This approach utilizes deep neural networks and explores various data modalities, including RG…
-
TEDDY foundation model predicts pediatric disease risk with high accuracy
Researchers have developed TEDDY, a novel foundation model designed to predict the risk of various diseases in children using historical diagnostic data. Trained on millions of ICD-10 diagnoses from over a million child…
-
Prostate MRI preprocessing boosts AI diagnostic accuracy for cancer detection
A new study published on arXiv investigates the impact of different diffusion-weighted imaging (DWI) preprocessing techniques on prostate MRI analysis. Researchers found that applying denoising, Gibbs-ringing correction…
-
New multimodal AI framework improves breast tumor classification accuracy
Researchers have developed a new multimodal framework for classifying breast fibroadenoma and phyllodes tumors, which often have overlapping appearances on ultrasound. This framework integrates visual, textual, and clin…
-
Student seeks advice on improving inconsistent diabetic retinopathy AI model
A computer engineering student is seeking advice on improving a 5-class diabetic retinopathy detection model trained on the APTOS 2019 dataset. The model exhibits inconsistent predictions, misclassifying classes like Mo…
-
New metric measures AI model robustness using Fisher Information
Researchers have developed a new method to measure the robustness of deep neural networks using the spectral norm of the Fisher Information Matrix (FIM). This attack-agnostic metric quantifies how sensitive a model's ou…
-
New Fisher Information metric assesses deep neural network robustness
Researchers have introduced a new metric for evaluating the robustness of deep neural networks, based on the spectral norm of the Fisher Information Matrix. This attack-agnostic approach offers theoretical bounds and pr…