Browsing by Author "Shen, Po-Chun"

Now showing 1 - 4 of 4

A Forward Collision Avoidance System Adopting Multi-feature Vehicle Detection
(2014-01-01)
Chen, Guan-Yu
;
Shen, Po-Chun
;
Cho, Chao-Yi
;
Vinay, M. S.
;
Guo, Jiun-In
This paper presents a vehicle collision avoidance method and it is associated with the exploitation of multiple features based on the AdaBoost classifier result. The proposed system will enhance the precision rate and accuracy of vehicle detection from 72% to 96% using an AdaBoost detection method extracting multiple features in vehicle detection and false alarms are reduced.
A Single-Camera High Dynamic Range Technique by Using Contrast Enhancement and Exposure Control
(2014)
Li, Che-Cheng
;
Hsu, Sheng-Wei
;
Shen, Po-Chun
;
Guo, Jiun-In
;
電子工程學系及電子研究所
This paper presents a single-camera HDR imaging algorithm, which is able to solve all the extreme conditions by controlling the exposure time to get a well-exposed image and then enhancing it. The proposed system is divided into two parts, i.e. single-image HDR imaging technique and an exposure control method. The single-image HDR imaging technique integrates two methods, i.e. Dynamic Local Contrast Enhancement (DLCE) and fuzzy enhancement. The integrated result can handle both glaring region and dark region to exhibit good visibility for extreme conditions. In exposure control method, we utilize local average, camera response function and main object detection to achieve a well-exposed image. Finally, we implement our system on TI OMAP4430 embedded platform. The optimized program is able to achieve about 13 frames per second with VGA resolution.
Dynamic Local Contrast Enhancement for Advanced Driver Assistance System in Harsh Environments
(2014-01-01)
Hsu, Sheng-Wei
;
Chen, Guan-Yu
;
Shen, Po-Chun
;
Cho, Chao-Yi
;
Guo, Jiun-In
This paper presents a Dynamic Local Contrast Enhancement (DLCE) method, which can strengthen the image quality in most of inclement weather conditions. This improves unnatural image over-enhancement, reduce noise, make image more saturated, and can be applied on foggy day and night. DLCE reaches 50 fps D1 resolution and 120 fps CIF resolution on ATOM 1.6GHz in TREK-668 embedded platform.
USING C3D TO DETECT REAR OVERTAKING BEHAVIOR
(2019-01-01)
Tseng, Ching-Kai
;
Liao, Chien-Chih
;
Shen, Po-Chun
;
Guo, Jiun-In
;
電子工程學系及電子研究所
Avoiding traffic accidents is critical since the death of traffic accidents is the eighth among the top ten leading causes of death in 2018. This paper proposes a light-weight convolutional 3D (C3D) network with five 3D convolution layers and two fully-connected layers to predict overtaking behavior. This network utilizes the last layer of convolution layer to learn the overtaking object location in the final frame. Based on NVIDIA Jetson TX2, the proposed C3D network achieves 91.46% accuracy to detect overtaking behavior on rainy days. To generate this excellent deep learning model, we use an efficient labeling tool, called ezLabel, which is a free SaaS for academia group with 96,000 opened image data samples for deep learning. ezLabel owns outstanding route prediction and fitting functions, which speeds up with the factor of ten compared to traditional tools. Users only label the object in its first frame and in its final frame, and then ezLabel labels the object in all frames in between and fits the bounding box to the object. The ezLabel can be used to label objects captured with any moving or static cameras efficiently.