A new framework called FuseDepth fuses frozen detection, matting and surface-normal models with an adapted Depth Anything backbone to achieve...

A new study finds that random forest models outperform neural networks in predicting arsenic contamination across Ghana's Pra, Ankobra, and...

Researchers at IIT Jodhpur have built a retrieval-augmented AI framework and a massive new dataset, FactViz3M, that can locate the...

A hybrid machine learning framework that reads leaf color, texture, and vein patterns diagnosed magnesium and iron deficiencies in field-grown...

Vietnamese researchers have trained a 3D U-Net to reconstruct facial soft tissue directly from skull geometry, achieving millimeter-level accuracy while...

A new graph neural network study of spatial transcriptomics finds that attention mechanisms, not elaborate edge features, drive accurate identification...

Researchers have introduced PMGL-GAD, a framework that combines training-free green learning features with a global attention disambiguation module to solve...

Researchers have built a WaveNet-CNN framework that fuses spatial and frequency-domain features with a dual moth flame optimization algorithm to...

Researchers at Tarbiat Modares University have developed DuoDiT, a dual-stream architecture that adapts large diffusion transformers for image generation by...

Researchers have developed a two-level evaluation framework showing that synthetic vertebrae which score highly on geometric similarity can still fail...