you can find code here: https://colab.research.google.com/drive/1lcpICRa4krlX7vcgjtnP8rU9Uw0WuA-r?authuser=1
- samples:
- thinned samples, for boundary recognition:
There are various methods for the recognition of handwritten digits such as pattern recognition, deep learning, or machine learning. In this project, we used the Sequence Matching algorithm to recognize the digits. This approach is used for comparison between unknown data and patterns. In order to calculate the distance between two images, initially, we convert them into signal and then recognise the boundary of the digits.The rationale is that the same handwritten digits have similar signals so that by the conversion of images into a single-dimensional sequence of numbers (signals), the algorithm was able to compare signals with some pattern signals and predict the label of images.
comparing two signals and calculating distances
results:
