Ali Wali

Video stabilization with moving object detecting and tracking for aerial video surveillance

Aerial surveillance system provides a large amount of data compared with traditional surveillance system. But, it usually suffers from undesired motion of cameras, which presents new challenges. These challenges must be overcome before such video can be widely used. In this paper, we present a novel video stabilization and moving object detection system based on camera motion estimation. We use local feature extraction and matching to estimate global motion and we demonstrate that Scale Invariant Feature Transform (SIFT) keypoints are suitable for the stabilization task. After estimating the global camera motion parameters using affine transformation, we detect moving object by Kalman filtering. For motion smoothing, we use a median filter to retain the desired motion. Finally, motion compensation is carried out to obtain a stabilized video sequence. A number of aerial video examples demonstrate the effectiveness of our proposed system. We use the software Virtual Dub with the Deshaker-Plugin for test purposes. For objective evaluation, we use Interframe Transformation Fidelity for video stabilization tasks and Detection Ratio for moving object detection task.

A New System for Event Detection from Video Surveillance Sequences

In this paper, we present an overview of a hybrid approach for event detection from video surveillance sequences that has been developed within the REGIMVid project. This system can be used to index and search the video sequence by the visual content. The platform provides moving object segmentation and tracking, High-level feature extraction and video event detection.We describe the architecture of the system as well as providing an overview of the descriptors supported to date. We then demonstrate the usefulness of the toolbox in the context of feature extraction, events learning and detection in large collection of video surveillance dataset.

Event Detection from Video Surveillance Data Based on Optical Flow Histogram and High-level Feature Extraction

This paper presents a new approach for event detection from video surveillance data based on optical ¿ow histogram with no prior knowledge of the motion nature. First,we start by estimating the motion from images sequence using optical flow technique. Second, we perform a classification using the histogram of the optical flow vectors and we use a chain coding algorithm that we applied to each class for the spatial segmentation. Finally, we extract a high-level feature from any frame for use in the learning and search events by SVM and HMM. We have tested the developed method on realimage sequences, our results are very promising.

Incremental Learning Approach for Events Detection from Large Video Dataset

In this paper, we propose a strategy of multi-SVM incrementallearning system based on Learn++ classifier for detectionof predefined events in the video. This strategy is offlineand fast in the sense that any new class of event can belearned by the system from very few examples. The extractionand synthesis of suitably video events are used for thispurpose. The results showed that the performance of oursystem is improving gradually and progressively as we increasethe number of such learning for each event. We thendemonstrate the usefulness of the toolbox in the context offeature extraction, concepts/events learning and detectionin large collection of video surveillance dataset.

Emotion recognition by analysis of EEG signals

We propose in this paper an emotional recognition system based on physiological signals. We adopt the seven basic emotions that are: neutrality, joy, sadness, fear, anger, disgust and surprise. An experiment has been conducted to verify the feasibility of the proposed system. This experience has allowed us to acquire EEG signals and to create an emotional database. For this, we have used the Emotiv EPOC headset. Thereafter, we have chosen the fuzzy logic techniques to classify the EEG signals and to analyze the results.

Analysis of Scintigraphic Renal Dynamic Studies: An Image Processing Tool for the Clinician and Researcher

The objective of this work is to develop a platform-independent tool for analysis of Scintigraphic renal dynamic studies. It allowing quantification kidney, from a series of Scintigraphic images in the format provided DICOM. This tool allows an automatic or manual drawing of regions of interest and the kidney in renal background even if small kidney and / or little functional kidney, drawing the isotopic nephrogram corrected for background noise (activity curves kidney) and determining the renal function on according to the method of the integral. This developed tool allows obtaining semi-automatic so reproducible results on page relevant information to the physician to assess the functional status of each kidney: the isotopic nephrogram, viewing dynamic images and the relative function that users can calculate renal function through the regions of both kidneys. This tool is a step forward towards standardization as a suitable tool for education, research, and for receiving distant expert’s opinions.

Multi-agent system for moving object segmentation and tracking

Video segmentation and tracking have been important and challenging issues for many video processing. A novel spatio-temporal video object segmentation and tracking algorithm is proposed in this paper. This algorithm is based on multi-agent system and active contour technique. The multi-agent system is composed of a set of supervisor and explorator agents. The agents are communicating and inspired in their conduct from active contour technique, more precisely the “Level Sets”. We used the DIMA platform to implement this algorithm. Experimental results indicate that the proposed algorithm is more robust than previous approaches.

An intelligent system for renal segmentation

Scintigraphic images are often characterized with much noise and a badly contrasted resolution which makes the perception of regions of interest very difficult. The renal quantification is how to define the regions of interests whose activities informs on the status of the renal function. In this context, the current study presents an intelligent system for the segmentation of renal regions in order to facilitate the process of quantification. The use of a multi-agent system based on the HOG3D descriptor combined with Fast Marching Method, has made our System of segmentation faster and more accurate. This automatic segmentation system is expected to assist physicians in both clinical diagnosis and educational training. Experiments and tests were developed on a database including 1800 images from 15 patients selected to obtain a variety of images. The results of the application of our method on several dynamic images are presented and discussed.

Towards an intelligent information research system based on the human behavior: Recognition of user emotional state

Our works deals with the problem of Information Retrieval System (IRS) that integrates the human behaviour. This system must be able to recognize the degree of satisfaction of the user of the result found through its facial expression, its physiological state, its gestures and its voice. For this, we propose in this paper an algorithm for recognizing the emotional state of a user during a search session in order to issue the relevant documents that he needs. We present also, the architecture agent of the envisaged system.

Moving Object Detection System in Aerial Video Surveillance

Moving object detection in aerial video, in which the camera is moving, is a complicated task. In this paper, we present a system to solve this problem by using scale invariant feature transform(SIFT) and Kalman Filter. Moving objects are detected by a feature point tracking method based on SIFT extraction and matching algorithm. In order to increase the precision of detection, some pre-processing methods are added to the surveillance system such as video stabilization and canny edge detection. Experimental results indicate that the suggested method of moving object detection can be achieved with a high detection ratio.