Post Processing Pedestrian Detection with Background Cues
ثبت نشده
چکیده
In this project, I develop an approach to post process local pedestrian detection by incorporating cues from the background. This approach uses a combination of several location specific color classifiers to quickly reduce the number of the local detector responses by discarding improbable results. An experimental method for color comparison is explored in this project to supplement existing techniques for color matching. Using the reduced subset feedback from the color classifiers, I estimate the horizon line position and partition the set into overlapping subsets. The final result is determined by the strength of each subset. This framework can be attached to any existing local pedestrian detector.
منابع مشابه
Combining Multiple Cues for Pedestrian Detection in Crowded Situations
This paper proposes a vision-based pedestrian detection in crowded situations based on a single camera. The main idea behind our work is to fuse multiple cues so that the major challenges, such as occlusion and complex background facing in the topic of crowd detection can be successfully overcome. Based on the assumption that human heads are visible, circle Hough transform (CHT) is applied to d...
متن کاملPedestrian detection by means of far-infrared stereo vision
This article presents a stereo system for the detection of pedestrians using far-infrared cameras. Since pedestrian detection in far-infrared images can be difficult in some environmental conditions, the system exploits three different detection approaches: warm area detection, edge-based detection, and disparity computation. A final validation process is performed using head morphological and ...
متن کاملAVSS2011 demo session: Real-time human detection using fast contour template matching for visual surveillance
Achieving accurate pedestrian detection for practically relevant scenarios in real-time is an important problem for many applications, while representing a major scientific challenge at the same time. We present a human detection framework which efficiently computes pedestrianspecific shape and motion cues and combines them in a probabilistic manner to infer the location and occlusion status of...
متن کاملLocal Associated Features for Pedestrian Detection
Local features are usually used to describe pedestrian appearance. While most of existing pedestrian detection methods don’t make full use of context cues, such as associated relationships between local different locations. This paper proposes two novel kinds of local associated features, gradient orientation associated feature (GOAF) and local difference of ACF (ACF-LD), to exploit context inf...
متن کاملA Trainable Pedestrian Detection System
In the near future, we can expect on-board automotive vision systems that inform or alert the driver about pedestrians, track surrounding vehicles, and read street signs. Object detection is fundamental to the success of this type of next-generation vision system. In this paper, we present a trainable object detection system that automatically learns to detect objects of a certain class in unco...
متن کامل