A comprehensive index of artificial intelligence and machine-learning research with AI-generated summaries, citation metrics, and direct links to papers and code.
Jie Hu, Li Shen, Samuel Albanie, et al.
The Squeeze-and-Excitation block adaptively recalibrates channel-wise feature responses by modelling interdependencies between channels, improving CNN performance at slight computational cost.
Chao Dong, Chen Change Loy, Kaiming He, et al.
This paper introduces SRCNN, a deep CNN that learns an end-to-end mapping for single image super-resolution, outperforming sparse-coding methods.
Shaoqing Ren, Kaiming He, Ross Girshick, et al.
Introduces a Region Proposal Network (RPN) that shares convolutional features with Fast R-CNN, enabling nearly cost-free region proposals and real-time object detection.
Unknown
A privacy-preserving FCNN framework using TFHE-based inference for secure medical image analysis on MedMNIST.
Yin Dai, Yifan Gao, Fayu Liu
TransMed combines CNN and transformer architectures for multi-modal medical image classification, achieving significant accuracy improvements on parotid gland and knee injury datasets.
Yipeng Hu, Marc Modat, Eli Gibson, et al.
This paper introduces a weakly-supervised CNN approach for multimodal image registration that learns voxel-level transformations from anatomical labels, enabling real-time, fully-automated inference without labels.
Unknown
SplineCNN introduces continuous B-spline kernels for geometric deep learning, enabling fast and flexible spatial aggregation on irregular data.
Unknown
This paper introduces a mixture model CNN framework for geometric deep learning on graphs and manifolds, achieving state-of-the-art results in deformable 3D shape analysis.
Unknown
This paper presents a unified geometric deep learning framework that generalizes CNNs and other architectures to non-Euclidean domains such as graphs and manifolds.
Marina Yusoff, Toto Haryanto, Heru Suhartanto, et al.
This structured review analyzes deep learning methods, especially CNNs and hybrids, for histopathological breast cancer image classification, highlighting challenges like overfitting and data imbalance.
Gabriel Eilertsen, Joel Kronander, György Dénes, et al.
This paper proposes a deep CNN specifically designed to predict lost information in saturated image areas, enabling HDR reconstruction from a single exposure.
Jo Schlemper, Ozan Oktay, Michiel Schaap, et al.
Introduces attention gates for CNNs that automatically learn to focus on salient regions in medical images, improving classification and segmentation performance without external localization modules.