Single-dataset orange image benchmark reports 99.12% accuracy with ViT, PCA and SVM
A cross-validation study of 1,090 images from one public orange dataset found 99.12% ± 0.71% accuracy for ViT + PCA ...
Tag: computer vision
Multimodal model grades plant disease severity from field images
AgriMamba combines field images with pathological descriptions to locate diseased leaves, segment lesions and grade severity, reporting competitive performance across ...
CTVSeg-UNet: robust refinement for leaf vein phenotyping in curing tobacco
CTVSeg-UNet improved vein-pixel recall from 47.12% to 77.03% over standard U-Net while extracting more continuous tobacco leaf networks across curing ...
Multimodal semantic segmentation measurement of shank length in live chickens based on SegFormer and geometric positioning strategy
A multimodal SegFormer system measured live-chicken shank length through anatomical masks and geometric anchors, achieving low repeated-measurement variation but substantial ...
Vision model estimates guidewire contact forces in vessel-phantom images
A multi-scale vision model estimated guidewire contact forces from planar images with a mean absolute error of 0.861 g on ...
Video-analysis study finds hybrid computer vision tracks soccer offside positions
A benchmark study combining adaptive background modeling, four-frame differencing and occlusion-aware tracking reached 92.9% mAP@0.5 on ISSIA-CNR and 93.1% on ...
Facial AI estimates biological age and predicts cancer survival
A foundation model trained on more than 40 million facial images produced tools that estimate biological age and directly predict ...
CDA-YOLO improves PCB defect detection over YOLOv11
An enhanced YOLOv11 model developed for printed circuit board inspection achieved mean average precision of 96.1% at an intersection-over-union threshold ...
AI Learns to Spot Family Resemblance by Fusing Siamese Networks with Vision Transformers
A hybrid deep learning framework combining Siamese networks and Vision Transformers achieves up to 95 percent accuracy in verifying parent-child ...
AI Learns Ancient Art Rules to Bring Fading Dunhuang Murals Back to Life
A new AI framework called GECST combines generative adversarial networks with classical compositional rules to restore China's fading Dunhuang murals ...
-
Mothers who receive childcare support from maternal grandparents show more parental warmth, finds NTU Singapore study
-
University of Seville Breaks 120-Year-Old Mystery, Revises a Key Einstein Concept
-
Bee body mass, pathogens and local climate influence heat tolerance
-
Researchers record first-ever images and data of a shark experiencing a boat strike
-
Groundbreaking Clinical Trial Reveals Lubiprostone Enhances Kidney Function
Embark on a thrilling journey of discovery with Scienmag.com—your ultimate source for cutting-edge breakthroughs. Immerse yourself in a world where curiosity knows no limits and tomorrow’s possibilities become today’s reality!