Research on human lightness perception has revealed important principles of how we perceive achromatic surface color, but has resulted in few image-computable models. Here we examine the performance of a recent artificial neural network architecture in a lightness matching task. We find similarities between the network’s behaviour and human perception. The network has human-like levels of partial lightness constancy, and its log reflectance matches are an approximately linear function of log illuminance, as is the case with human observers. We also find that previous computational models of lightness perception have much weaker lightness constancy than is typical of human observers. We briefly discuss some challenges and possible future directions for using artificial neural networks as a starting point for models of human lightness perception.
In future manufacturing human-machine interaction will evolve towards flexible and smart collaboration. It will meet requirements from the optimization of assembly processes as well as from motivated and skilled human behavior. Recently, human factors engineering has substantially progressed by means of detailed task analysis. However, there is still a lack in precise measuring cognitive and sensorimotor patterns for the analysis of long-term mental and physical strain. This work presents a novel methodology that enables real-time measurement of cognitive load based on executive function analyses as well as biomechanical strain from non-obtrusive wearable sensors. The methodology works on 3D information recovery of the working cell using a precise stereo measurement device. The worker is equipped with eye tracking glasses and a set of wearable accelerometers. Wireless connectivity transmits the sensor-based data to a nearby PC for monitoring. Data analytics then recovers the 3D geometry of gaze and viewing frustum within the working cell and furthermore extracts the worker's task switching rate as well as a skeleton-based approximation of worker's posture associated with an estimation of biomechanical strain of muscles and joints. First results enhanced by AI-based estimators demonstrate a good match with the results of an activity analysis performed by occupational therapists.
Cyber security has become an increasingly important topic in recent years. The increasing popularity of systems and devices such as computers, servers, smartphones, tablets and smart home devices is causing a rapidly increasing attack surface. In addition, there are a variety of security vulnerabilities in software and hardware that make the security situation more complex and unclear. Many of these systems and devices also process personal or secret data and control critical processes in the industry. The need for security is tremendously high. The owners and administrators of modern computer systems are often overwhelmed with the task of securing their systems as the systems become more complex and the attack methods increasingly intelligent. In these days a there are a lot of encryption and hiding techniques available. They are used to make the detection of malicious software with signature based scanning methods very difficult. Therefore, novel methods for the detection of such threats are necessary. This paper examines whether cyber threats can be detected using modern artificial intelligence methods. We develop, describe and test a prototype for windows systems based on neural networks. In particular, an anomaly detection based on autoencoders is used. As this approach has shown, it is possible to detect a wide range of threats using artificial intelligence. Based on the approach in this work, this research topic should be continued to be investigated. Especially cloud-based solutions based on this principle seem to be very promising to protect against modern threats in the world of cyber security.