Visual learning is expected to be a continuous and robust process, which treats input images and pixels selectively. In this paper we present a method for subspace learning, which...
Support Vector Machines (SVMs) have been successfully applied to solve a large number of classification and regression problems. However, SVMs suffer from the catastrophic forgetti...
This paper describes a new hybrid architecture for an artificial neural network classifier that enables incremental learning. The learning algorithm of the proposed architecture d...
Abstract— This paper presents i-AA1 , a constructive, incremental learning algorithm for a special class of weightless, self-organizing networks. In i-AA1 , learning consists of ...
Hand-coded vision systems are problematic in complex medical domains and are hard to change as new information emerges. Knowledge-Engineering and Machine Learning approaches to in...