Sharif Digital Repository

The purpose of multi target tracking, is tracking multiple target in each frame of video sequences simultaneously. Object tracking encounters many challenges. Two major challenges of tracking problem that we focus on in them in this thesis are detecting object in each frame of video sequence and handling occlusion problem. Our proposed method for multi target tracking considers one target as a main target and other targets as subsidiary targets. Proposed method track objects in initial view by Kalman filter that is adapted and tuned for multi target tracking. In case of occlusion our method uses data from auxiliary view for handling occlusion. In auxiliary view objects that are occluded in... 

Today, advances in technology caused the emergence of a need for intelligent systems which can operate without human supervision, and have variety of applications in security, industry, transportation and sport events. Most of these systems require faster and more accurate object tracking. Lots of surveillance and computer vision applications need a process to detect and track objects. For example, the exact location of objects or people is part of the input for applications like movement analysis, counting the number of items, and recognition of people’s behavior. Occlusion is one of the major obstacles in performance of auto-tracking systems. There are several methods to address this... 

Image classification is one of the most challenging issues in computer vision. One sort of such classifications is Scene Classification. To perform automatic classification reserchers used many aproches.The general approach used features directly extracted from the image, such as color and texture or features extracted by the SIFT algorithmetc. Another method is based on recognizing object of the Scene (espessially indoor scene). This method is based on finding of a limited number of prespecified objects. In the proposed method, first a window surrounding each objects, (regardless of the type of object) founded. Then the SIFT feature is extracted from that window. All features (corresponding... 

Today, due to the increasing activity of terrorist groups, monitoring people in important and busy places such as airports and train stations is very important. One of the technologies that has been developed for this purpose in recent years is 3D imaging technology using millimeter wave. These systems use millimeter waves to image people and identify objects hidden under clothing, which do not have the limitations of conventional imaging techniques such as x-rays and metal detectors. One of the advantages of using these systems is the ability to automatically detect objects in millimeter wave images using deep neural networks such as Segmented, Faster R-CNN and Mask R-CNN, which using these... 

Multiple object tracking refers to the detection and following of target object classes in video sequences. In this task, all objects belonging to the target classes in the video are detected simultaneously in each frame, and a unique ID is assigned to each of them throughout the video. In recent years, the use of graph neural networks for solving this problem has received significant attention because these models are suitable tools for discovering and improving the relationships between objects in the scene, which can greatly assist in better object pairing. However, there are various challenges to using graph neural networks, the most important of which is the limitation of input graph... 

Detection and Tracking objects in various environmental conditions is a challenging task in computer vision. Abrupt changes in illumination, object size and noise level make this task even harder. Due to these problems, it is somehow impossible to propose a fully functional system for tracking every type of object. In this thesis we propose a new method for detection and tracking vehicles in traffic scenes. The problem is solved by dividing it into two parts: at First we have evaluated more than ten state-of-the art trackers, the final multi-vehicle tracker is chosen, and then, like any other object tracking method, we have used a unique identiy for each vehicle. This traffic identity is... 

The existing learning methods usually assume that training and test data follow the same distribution, while this is not always true. Thus, in many cases the performance of these learning methods on the test data will be severely degraded. We often have sufficient labeled training data from a source domain but wish to learn a classifier which performs well on a target domain with a different distribution and no labeled training data. In this thesis, we study the problem of unsupervised domain adaptation, where no labeled data in the target domain is available. We propose a framework which finds a new representation for both the source and the target domain in which the distance between these... 

This project aims to highlight vision related tasks centered around "car". First, we gathered a dataset of 4343 front view car images, captured from the streets of Iran and Syria during daylight, the images of which are all manually cropped around their corresponding accurately chosen bounding boxes. we also extracted seven parts (i.e. left and right front lights, left and right mirrors, bumper, plate, and air intake) from each car image in the dataset. Our dataset is suitable for developing and testing bounding box extraction algorithms, holistic and part based analyses, occlusion handling algorithms, etc. next, we utilized Viola-Jones Detector to develop a system for car detection, in...