Comments (2)
有以下几个问题向您请教:
1、我复现出来的ShanghaiTech结果只有96%左右,没有修改任何代码,不知道什么原因;
2、crop的次数对识别率的影响特别大,我从测试集中的10次crop中随机取出一次crop来测试,识别率降到了88%左右;
3、和其他论文比较的识别率的时候,RTFM是做过10crop来训练和测试的,但是请问其他论文也做过10次crop吗?如果没有的话,那就有失公平了。
4、在Nvidia 2080Ti的inference time=0.76s,请问这个时间是10次crop的总时间还是单次crop的时间?
5、由于RTFM采用的多尺度技术,因此对于一段视频,如果有N个16帧的clips,最好的方式是这个N个clip全部输入到网络里识别是最好的,也正如测试代码所示,clips越少识别率越低。如果视频中的的N个clips单独测试,识别率也非常低,会降到80%左右。那么在实时监控的时候,每次就应该输入32个clips才能达到可以接受的识别率,但是这种方式,会导致实时性严重下降,甚至不可用。
there might be some auc differences between different computers and different runs.
In the previous papers, 'Not only Look, but also Listen: Learning Multimodal Violence Detection under Weak Supervision' and
'Graph Convolutional Label Noise Cleaner: Train a Plug-and-play Action Classifier for Anomaly Detection'. They all use 10-crop. Some of the other papers did not provide code. So I'm not sure with them though...
The time should be the one for single-cropped images. For question 5, I agree that it will harm some of the efficiency of the method when use more clips as input.
from rtfm.
@tianyu0207 Thanks!
from rtfm.
Related Issues (20)
- some questions about extracting I3D features
- xd-violence
- How to implement inference code? HOT 1
- Where do I find dir test_frame_mask, or make it? HOT 1
- how to make ground truth file and list file for shanghai tech dataset
- UCSD-Ped2 实验设置
- Custom .mat files HOT 3
- Can not reproduce the provided I3D features extracted from UCF-Crime on RTFM Github
- The UCF-Crime checkpoints link is disabled HOT 2
- Only ShanghaiTech features and checkpoint for ShanghaiTech can be used in the link
- UCF-Crime train I3d features on Google drive 下载失败—已被禁止 HOT 1
- Feature extraction results did not meet expectations
- feature extraction
- Results can't be reproduce for Sanghaitech
- Feature Extraction
- Generation of MatLab Labels
- AUC for UCF-Crime is Too low and Can't reproducible. HOT 1
- Can't download UCF_Crime test dataset
- 怎么生成ucf-i3d.list和ucf-i3d-test.list?
- checkpoints for UCF-Crime and XD-Violence datasets
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from rtfm.