Signature Verification Project
Collection
Using CycleGAN, YOLOv5 and VGG16 transfer learned on custom signature datasets to check for forging of signatures.
•
5 items
•
Updated
CNN based on VGG16 with bounding box regression layer -> A pre existing CNN architecture VGG16 was used. The final fully connected layer was removed and replaced with a bounding box regression layer. The dataset was the 776 images (subset of tobacco800) which had signatures. This model was used to predict the bounding box coordinates of signatures on documents.