The goals in the framework of the project
The CreativeLab is a project to develop a computer vision system with the aim to measure and process real-time data from individual activity in sports context. This project emerged from a holistic understanding of science and aims to develop a valid methodology to produce results that can contribute to the understanding of the complexity of human mind, throughout the exploration of collaborative work and creativity concepts.
Methodologies and approaches proposed to reach the goalsThe CreativeLab is a project focused in developing a system to measure performance indicators related to collaborative work and creativity, helping to explain the mechanisms that regulate the decision-making process and contribute to understanding the complexity of human mind. The new technological developments in computer vision domain make possible to measure and process, in real time, data related to the dynamical positioning of each subject when captured at high frequencies (70 images per second) and synchronize them on a timeline with the descriptors of the actions performed through notational analysis. These ideas allow overcoming the problems in current research, which is poor and scarce because it is limited to the use of questionnaires and laboratory-based testing. Indeed, the measures of creativity related indicators seem to imply the use of non-invasive methodologies that replicate reliably the environment of his natural expression. The possibility of developing a structure like CreativeLab, with the aim of producing performance results that allow evaluating the subjects in their collective environment in real time is a great advance in science. Also, the possibility to clarify and potentiate the transfer effects between the sport system and the labor and educational systems are of undeniable importance. Similarly, the possibility of extending the development of the system to the measure of teacher activity in order to improve its intervention is also very necessary, especially to less experienced teachers.
The development of CreativeLab will be done in three different stages prior to his full usage. The first stage is technical and comprehends the adaptation of a sports court located in the University of Trás-os-Montes and Alto Douro. This stage will be carried by placing and fixing the video cameras and by placing the wiring system to obtain a full-court coverage. It also comprehends the physical installation of the main computer that will receive and process all data from the cameras.
The second step comprises the development of data processing software with the appropriate algorithms and the inclusion of the collaborative work and creativity behavior indicators. The system is based in new developments in computer vision, through the use of four synchronized video cameras that record the dynamical positions of each subject at high frequencies. The problem of tracking an object in a sequence of images will be addressed with the ultimate goal of tracing the objects movement and behavior. Robust tracking methodologies are still a challenge because several difficulties can arise due to noise presence in images, quick changes in lighting conditions, abrupt or complex object motion, changing appearance patterns of the object and the scene, non-rigid object structures, object-to-object and object-to-scene occlusions and real time processing requirements. The issued tracking concept refers to the process of locating moving objects (players and ball) and estimating their trajectory along a sequence of images. In its simplest approach, tracking involves two steps: the detection of interest objects and the tracing of their trajectories. Furthermore, this work will also involve individual players identification, behavior analysis and activity classification. Behavior analysis and activity classification will be further interpretations of the tracking results. The tracking process will consist of two main tasks: figure-ground segmentation and temporal correspondences. Figure-ground segmentation is the process of extracting the objects or features of interest from the video sequence, i.e., the players and ball. At this step, segmentation methods based either on motion, appearance, shape, depth will be developed. Temporal correspondence concerns the association of the detected objects in the current frame with those in the previous frames defining temporal trajectories. In both tasks, fuzzy logic concepts will be used in order to endow the process with the needed robustness against the noise and uncertainty inherent to these kinds of problems.
The third stage is for system testing and inspection of validity and reliability. In this particular area, scientific papers will be produced to validate the measures of individual distance and velocity of the subjects in sports activity. Also, an inspection of reliability from the notational analysis data in real-time will be performed.
These previous steps will be leaded by different elements of the proposed research team. This team is composed by an electrical engineer specialized in the development of computer vision systems, one specialist research methods applied to social sciences to design and carry out the validation studies, one specialist in data processing using non-linear techniques, two specialists in decision-making and sports behavior, one specialist in pedagogical supervision and one specialist in creativity and collaborative work. The CreativeLab also has the collaboration from the University of Vienna, who has in course a similar project using radiofrequency and invasive sensors (Match and competitions analysis with the help of the realtime positioning system Ubisense, PID-ID:101620). Additionally, it has the collaboration from the University of Vigo, who has experience in developing projects in real-time notational analysis systems, similar to those projected for CreativeLab (Modeling competition in team sports by using computer tablets, DEP2011-23338).
After this development period, the system will be ready to fully operate and some pilot studies are already projected, specifically, to describe and compare the nature of interactions among subjects; to describe the effects of manipulating the environment information in the collective decision-making process; to contrast the subjects’ profiles throughout collective variables by describing their relative motion and developing a dynamical model of spatial behavior that includes different effects of subjects’ collaborative work and creativity.