Dynamic Computational Time for Visual Attention

March 30, 2017 · Entered Twilight · 🏛 2017 IEEE International Conference on Computer Vision Workshops (ICCVW)

"Last commit was 8.0 years ago (≥5 year threshold)"

Evidence collected by the PWNC Scanner

Repo contents: README.md, RecurrentAttention.lua, RecurrentAttentionDynamic.lua, SpatialGlimpseCrop.lua, VRClassRewardN.lua, checkpoints.lua, dataloader.lua, datasets, demo.sh, main.lua, mnist, models, opts.lua, process.py, save, train.lua, utils, weight-init.lua

Authors Zhichao Li, Yi Yang, Xiao Liu, Feng Zhou, Shilei Wen, Wei Xu arXiv ID 1703.10332 Category cs.CV: Computer Vision Citations 117 Venue 2017 IEEE International Conference on Computer Vision Workshops (ICCVW) Repository https://github.com/baidu-research/DT-RAM ⭐ 64 Last Checked 1 month ago

Abstract

We propose a dynamic computational time model to accelerate the average processing time for recurrent visual attention (RAM). Rather than attention with a fixed number of steps for each input image, the model learns to decide when to stop on the fly. To achieve this, we add an additional continue/stop action per time step to RAM and use reinforcement learning to learn both the optimal attention policy and stopping policy. The modification is simple but could dramatically save the average computational time while keeping the same recognition performance as RAM. Experimental results on CUB-200-2011 and Stanford Cars dataset demonstrate the dynamic computational model can work effectively for fine-grained image recognition.The source code of this paper can be obtained from https://github.com/baidu-research/DT-RAM