Video Object Detection with an Aligned Spatial-Temporal Memory

December 18, 2017 · Entered Twilight · 🏛 European Conference on Computer Vision

"No code URL or promise found in abstract"
"Code repo scraped from project page (backfill)"

Evidence collected by the PWNC Scanner

Repo contents: .gitignore, BatchProviderBase_StaticImg.lua, BatchProviderROI_StaticImg.lua, BatchProviderVID.lua, ColorTransformer.lua, DataSetJSON.lua, ImageDetect.lua, LICENSE, Makefile, README.md, Tester_VID.lua, config.lua, data_static.lua, data_video.lua, donkey_static.lua, engines, external, fbcoco.lua, imgs, loaders, models, modules, myutils.lua, nms.c, scripts, train_video.lua, utils.lua

Authors Fanyi Xiao, Yong Jae Lee arXiv ID 1712.06317 Category cs.CV: Computer Vision Citations 199 Venue European Conference on Computer Vision Repository https://github.com/fanyix/STMN ⭐ 121 Last Checked 6 days ago

Abstract

We introduce Spatial-Temporal Memory Networks for video object detection. At its core, a novel Spatial-Temporal Memory module (STMM) serves as the recurrent computation unit to model long-term temporal appearance and motion dynamics. The STMM's design enables full integration of pretrained backbone CNN weights, which we find to be critical for accurate detection. Furthermore, in order to tackle object motion in videos, we propose a novel MatchTrans module to align the spatial-temporal memory from frame to frame. Our method produces state-of-the-art results on the benchmark ImageNet VID dataset, and our ablative studies clearly demonstrate the contribution of our different design choices. We release our code and models at http://fanyix.cs.ucdavis.edu/project/stmn/project.html.