Transcription of Tech report (v5) - arXiv
{{id}} {{{paragraph}}}
Example: barber
Tech report (v5) - arXiv
Rich feature hierarchies for accurate object detection and semantic segmentationTech report (v5)Ross Girshick Jeff Donahue Trevor Darrell Jitendra MalikUC detection performance, as measured on thecanonical PASCAL VOC dataset, has plateaued in the lastfew years. The best-performing methods are complex en-semble systems that typically combine multiple low-levelimage features with high-level context. In this paper, wepropose a simple and scalable detection algorithm that im-proves mean average precision (mAP) by more than 30%relative to the previous best result on VOC 2012 achievinga mAP of Our approach combines two key insights:(1) one can apply high-capacity convolutional neural net-works (CNNs) to bottom-up region proposals in order tolocalize and segment objects and (2) when labeled trainingdata is scarce, supervised pre-training for an auxiliary task,followed by domain-specific fine-tuning, yields a significantperformance boost.
image classification and object detection. This paper is the first to show that a CNN can lead to dramatically higher ob-ject detection performance on PASCAL VOC as compared to systems based on simpler HOG-like features. To achieve this result, we focused on two problems: localizing objects with a deep network and training a high-capacity model
Loading..
Tags:
Information
Domain:
Source:
Link to this page:
Please notify us if you found a problem with this document: