Abstract Traffic scene object detection and recognition is extensively researched in the field of roadside assistance. Due to its importance, many methods have been proposed to solve the classification of objects in traffic and aim classification in different lighting conditions, scaling, orientation and shape of objects. Although most methods for classification are binary classification,…
Abstract This paper presents the first results obtained by exploring different neighborhoods in two-dimensional Cellular Automata applied for the difficult task of automatic image segmentation. Numerical experiments have been performed on several real-world and synthetic images for which the ground truth is known, being therefore able to compute the algorithm performance by comparing the…
Abstract Cellular Automata (CA) can be successfully applied to the task of image segmentation. The CA-based GrowCut algorithm is able to perform such a task and we aim to investigate the full emergence phenomenon that arises during the segmentation process. In fact, we want to investigate how the segmentation performance could depend on the…
Abstract Multiclass classification is an extensively researched topic due to its importance in making the binary classification problems a complex and well tuned system and minimising the running time for multiple classification problems. In the traffic scenes one can encounter several types of obstacles like cars, pedestrians, animals, low elevated objects, road signs that…
R. Mocan, L. Diosan, “Support vector machine and boosting based Multiclass classification for traffic scene obstacles”, Studia Univ. Babes–Bolyai, Informatica, Vol. 61, Nr 2, 2016
R. Mocan, L. Diosan, “Multiclass classification based on clustering approaches for obstacle recognition in traffic scenes”, 2016 IEEE 12th International Conference on Intelligent Computer Communication and Processing (ICCP), 2016, pp. 257 – 261
Abstract It was proved that the fusion of information from multi-modality images increases the accuracy of pedestrian recognition systems. One of the best approach so far is to concatenate the features from multi-modality images into a large feature vector, but it requires strong camera calibration settings and non-discriminative modalities could lead to missclassification of…
Abstract Despite many years of research, pedestrian recognition is still a difficult, but very important task. It was proved that concatenating information from multi-modality images improves the recognition accuracy, but with a high computational cost. We present a modality selection approach, which is able to dynamically select the most discriminative modality for a given…
Abstract The detection of tumors in digital images originated from mammograms can be a challenging task. In this paper we investigate a Computer Aided Diagnosis System based on a Genetic Programming classifier. The performance of the considered classifier is evaluated for five of the image descriptors used in literature and we propose a new…
Abstract Despite many years of research, breast cancer detection is still a difficult, but very important problem to be solved. An automatic diagnosis system could establish whether a mammography presents tumours or belongs to a healthy patient and could offer, in this way, a second opinion to a radiologist that tries to establish a…