swMATH ID: 28335
Software Authors: Kye-Hyeon Kim, Sanghoon Hong, Byungseok Roh, Yeongjae Cheon, Minje Park
Description: PVANET: Deep but Lightweight Neural Networks for Real-time Object Detection. This paper presents how we can achieve the state-of-the-art accuracy in multi-category object detection task while minimizing the computational cost by adapting and combining recent technical innovations. Following the common pipeline of ”CNN feature extraction + region proposal + RoI classification”, we mainly redesign the feature extraction part, since region proposal part is not computationally expensive and classification part can be efficiently compressed with common techniques like truncated SVD. Our design principle is ”less channels with more layers” and adoption of some building blocks including concatenated ReLU, Inception, and HyperNet. The designed network is deep and thin and trained with the help of batch normalization, residual connections, and learning rate scheduling based on plateau detection. We obtained solid results on well-known object detection benchmarks: 83.8
Homepage: https://arxiv.org/abs/1608.08021
Source Code: https://github.com/sanghoon/pva-faster-rcnn
Related Software: Caffe; ImageNet; AlexNet; XNOR-Net; Inception-v4; DoReFa-Net; SqueezeNet; MobileNets; MS-COCO; ShuffleNet; Faster R-CNN; SSD; fbfft; cuda-convnet; LCNN; TensorFlow; Xception; DOTA; LabelMe; PASCAL VOC
Cited in: 3 Publications

Citations by Year