Stamp and Signature Detector
The Application requires training for the stamps and signatures that are to be detected. To train the application, at least 150 images for each stamp or signature are required
Modern object detectors and classifiers require heavy computing power to train the models as well as run the models for testing. Our goal was to build such a model that is light weight, requires less computing power comparatively, and provides best results.
The Stamp and Signature Detector application consists of combination of Support Vector Machine (SVM) and Histogram of Gradient (HOG). SVM is a supervised learning model with associated learning algorithm that analyze data for classification and regression analysis. HOG is a feature descriptor, which has the capability to extract features using CPU only.
The Application requires training for the stamps and signatures that are to be detected. To train the application, at least 150 images for each stamp or signature are required
Technology Stack:
- Python 3.7
- OpenCV – HOG
- Dlib – SVM
Benefits:
- Verify collection of Documents in an instant
- Faster than Manual Verification
- Usable with Lightweight Hardware
- Usable as a part in Process Automation