Image recognition with machine learning: an adaptive approach

Image recognition is a recurring theme in the field of computer vision and it currently borrows some elements from Artificial Intelligence also. However, images are often dealt with in a specific manner, with the algorithms exploring some feature of the said image (be it during pre-processing or during image recognition). Such as is the case when, for instance, some specific algorithms are used for recognition purposes (ranging from biometric recognition to medical images recognition and also the recognition of some defects in electrical / electronic circuit boards), to name just a few. In this article, we propose instead an adaptive algorithm for image recognition, in such a way that an image is treated as if it were composed of a set of pixels - without any other specific parameter. Simple Artificial Intelligence elements are employed in this approach, such as the learning algorithm and the cluster concept.