Robust poker image recognition scheme in playing card machine using Hotelling transform, DCT and run-length techniques

Poker is an interesting field for artificial intelligence research. It is a game of imperfect information and chance associated outcomes, where players deal with probability, risk assessment, and possible deception – just like real life decision-making. In our proposal, three strategies are used to organize a robust poker recognition scheme: (1) using Hotelling transform to place the object image in the correct position in poker pick-up stage; (2) a weighted compacted energy (WCE) of the image is used as the first feature in using DWT and DCT; and (3) calculating four orientation connectivity run-length values (FOCRLV) to distinguish different poker card images. There are two contributions in this article – one is the use of FOCRLV as a special feature to improve image recognition and the other is use of the compact energy band of an image as another feature to effectively identify the images. In order to demonstrate the effectiveness of the proposed scheme, simulations under various conditions were conducted. The experimental results show that our proposed scheme can exactly identify ranks and suits of the poker images 100% of the time, even when 40% noise is added, or when intensity level is increased or decreased 40%.