The cost of drawing object bounding boxes (i.e. labeling) for millions of images is prohibitively high. For instance, labeling pedestrians in a regular urban image could take 35 seconds on average. Active learning aims to reduce the cost of labeling by selecting only those images that are informative to improve the detection network accuracy. In this paper, we propose a method to perform active learning of object detectors based on convolutional neural networks. We propose a new image-level scoring process to rank unlabeled images for their automatic selection, which clearly outperforms classical scores. The proposed method can be applied to videos and sets of still images. In the former case, temporal selection rules can complement our scoring process. As a relevant use case, we extensively study the performance of our method on the task of pedestrian detection. Overall, the experiments show that the proposed method performs better than random selection.
Introduction
Medical education is changing and evolving. Teachers need to re-evaluate their medical teaching practice to enhance student learning. The data about the ideal training method of Basic Life Support (BLS) is lacking. The goal of this study was to analyse the use and performance of video self-instruction (VSI) method in BLS, in order to develop an efficient BLS training method.
Methods
Eighty-one undergraduate medical interns were enrolled in a prospective clinical study in 2011. They were divided into VSI group and traditional group. We provided the first group with a DVD containing a 20-minute training video while the second group took part in a 4-hour training class of BLS. Subjects participated in a pre-test and post-test based on 2010 American Heart Association Resuscitation guideline.
Results
The average scores of VSI group and the traditional group before training were 8.85±2.42 and 8.57±2.22 respectively (p=0.592). After training, the average scores of the VSI and the traditional group were 20.24±0.83 and 18.05±1.86 respectively. VSI group achieved slightly better scores compared with the traditional group (p<0.001).
Conclusions
Training through VSI achieves more satisfying results than the traditional lecture method. VSI method can be considered a useful technique in undergraduate educational programs. Developing VSI can increase significantly the access to the BLS training. (Hong Kong j.emerg.med. 2015;22:291-296)
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.