Search results
(1 - 3 of 3)
- Title
- Multi-Resolution Depth Estimation Using Modified Census Transform for Advanced Driver Assistance Systems
- Creator
- Loghman, Maziar, Mesmakhosroshahi, Maral, Kim, Joohee
- Date
- 2018-01-04, 2017-09-05
- Description
-
A computer-implemented depth estimation method based on non-parametric Census transform with adaptive window patterns and semi-global...
Show moreA computer-implemented depth estimation method based on non-parametric Census transform with adaptive window patterns and semi-global optimization. A modified cross-based cost aggregation technique adaptively creates the shape of the cross for each pixel distinctly. In addition, a depth refinement algorithm fills holes within the estimated depth map using the surrounding background depth pixels and sharpens the object boundaries by exerting a trilateral filter to the generated depth map. The trilateral filter uses the curvature of pixels as well as texture and depth information to sharpen the edges.
Sponsorship: Illinois Institute of Technology
United States Patent
Show less
- Title
- PEDESTRIAN DETECTION AND TRACKING FOR ADVANCED DRIVER ASSISTANCE SYSTEMS
- Creator
- Mesmakhosroshahi, Maral
- Date
- 2017, 2017-05
- Description
-
In an effort to reduce driver errors in being the major cause of traffic accidents, there is a lot of research being conducted into the...
Show moreIn an effort to reduce driver errors in being the major cause of traffic accidents, there is a lot of research being conducted into the development of advanced driver assistance systems (ADAS). ADAS is a system aimed at helping the driver in tasks such as pedestrian and vehicle detection, traffic sign recognition and lane detection. Pedestrian detection is one of the major tasks in advanced driver assistance systems (ADAS). Most of the stereo based pedestrian detection algorithms include three major steps: 1. Ground plane estimation 2. Region of interest (ROI) generation 3. Pedestrian classification In this thesis, we present a stereo-based pedestrian detection framework for advanced driver assistance systems by exploiting both color and depth information obtained from a stereo camera installed in a vehicle. In our proposed framework, we first use the vertical gradient of the dense depth map to estimate and discard the ground plane. The boundaries of the ground plane are then searched to detect the pedestrians and the depth values of the boundaries are used to compute the size of the detection windows for detecting pedestrians at different scales. In addition, a depth-based multi-scale ROI extraction method has been proposed to reduce the computation time of ROI extraction. For classifying ROIs to pedestrian and non-pedestrian classes, Histogram of Oriented Gradients (HOG)/Linear support vector machine (SVM) and Integral Channel Features (ICF)/Adaboost are used. To recover the missed pedestrians and improve the detection rate, an ROI tracking algorithm is proposed which incorporates the ROIs extracted from the current frame with theROIs tracked from a reference frame. For additional reduction in search space, we propose a novel algorithm to reduce the number of candidate windows extracted as ROI by taking advantage of the temporal correlation between the adjacent frames. We also propose a method to improve the accuracy of the pedestrian classifi- cation using the aggregated channel features. In this approach, we use the aggregated channel features as our baseline detector and improve it's accuracy using the tanh normalization and Gabor filter. After classification using Adaboost, we use a posi- tive subset of the bounding boxes to classify them again using Convlutional Neural Network to finalize the detection.
Ph.D. in Electrical Engineering, May 2017
Show less
- Title
- ACTION RECOGNITION USING SPATIO-TEMPORAL FEATURE EXTRACTION
- Creator
- Mesmakhosroshahi, Maral
- Date
- 2012-04-24, 2012-05
- Description
-
The importance of automatically recognition of human activities and it's ap- plications increase as the technology is growing fast these days....
Show moreThe importance of automatically recognition of human activities and it's ap- plications increase as the technology is growing fast these days. In this thesis we focus on the action recognition techniques using 3D feature extraction techniques and try to improve the performance of the available feature extraction methods to have more accurate action classi cation. This thesis is based on spatio-temporal feature extraction techniques that can be extension of two dimensional feature extractions. At rst, 3D corner detector and 3D blob detector, two popular interest point detectors that successfully extended to spatio-temporal domain are reviewed. 3D corner detector is an extension of Harris corner detector which is proposed for corner detection in space-time and 3D blob detector is the 3D version of Hessian blob detector. A method is proposed to improve the robustness of 3D corner detectors against illumination variances based on the application of sigmoid function in contrast stretching. These interest points need to be described by feature vectors. In this part also a review on the SIFT descriptor and it's 3D extension is done. Another 3D feature descriptor discussed in this thesis is Cuboid. In the feature description part a new method is proposed based on the 3D SIFT feature descriptor by changing the binning method of the gradient orientation histogram to a non-uniform binning scheme. Using these non-uniform bins increases the accuracy of action classi cation by focusing on the areas near the interest points. Action classi cation is done using bag of features and supervised learning techniques. For making a dictionary of features, feature vectors are clustered by K-means clustering method and support vector machine is used for classi cation of human activities.
M.S. in Electrical Engineering, May 2012
Show less