- 标题
- 摘要
- 关键词
- 实验方案
- 产品
-
Free-Space Detection with Self-Supervised and Online Trained Fully Convolutional Networks
摘要： Recently, vision-based Advanced Driver Assist Systems have gained broad interest. In this work, we investigate free-space detection, for which we propose to employ a Fully Convolutional Network (FCN). We show that this FCN can be trained in a self-supervised manner and achieve similar results compared to training on manually annotated data, thereby reducing the need for large manually annotated training sets. To this end, our self-supervised training relies on a stereo-vision disparity system, to automatically generate (weak) training labels for the color-based FCN. Additionally, our self-supervised training facilitates online training of the FCN instead of offline. Consequently, given that the applied FCN is relatively small, the free-space analysis becomes highly adaptive to any traffic scene that the vehicle encounters. We have validated our algorithm using publicly available data and on a new challenging benchmark dataset that is released with this paper. Experiments show that the online training boosts performance with 5% when compared to offline training, both for Fmax and AP.
关键词： Self-supervised training,Fully Convolutional Network,Advanced Driver Assist Systems,Online training,Free-space detection
更新于2025-09-10 09:29:36
-
[IEEE 2018 1st International Cognitive Cities Conference (IC3) - Okinawa, Japan (2018.8.7-2018.8.9)] 2018 1st International Cognitive Cities Conference (IC3) - Hand Gesture Recognition Using Color-Depth Association for Smart Home
摘要： This study we propose a robust hand gesture segmentation method which associates the depth and color information with online training. Different existing methods, when the hands close to the body part or in cluttered background, our system remains valid. In the proposed method, a judging procedure of the hand point is applied to obtain more accurate result. As Kinect sensor has relatively large errors in acquisition of depth data at the edge of the object, this will lead to the wrong results with the depth threshold in coarse segmentation step. To compare with the existing systems, the proposed method is the least restrictive one in practically.
关键词： online training,Kinect sensor,Depth,ellipse model,hand gesture segmentation
更新于2025-09-09 09:28:46