zhanghengdev/gaff

[WACV 2021]"Guided Attentive Feature Fusion for Multispectral Pedestrian Detection"

gaff's Issues

求源码

你好，我是北京航空航天大学的的一名本科生，研究方向是红外可见光行人检测，用于本科毕设，不做商业用途。我想请问一下，您这个模型的代码能否开源或者发给我呢？我的邮箱是[email protected]。万分感谢！！

关于开源

你好，我是陆军工程大学的一名硕士研究生，研究方向是模型轻量化。我目前的大论文是做红外目标检测模型的轻量化，我想请问一下，您这个模型的代码能否开源或者发给我呢？我的邮箱是[email protected]。万分感谢！！

The work is so interesting,do you plan to open the great source,please!

Hi,thank you for your great work!
Do you plan to open the source?
The work is so interesting,hope the source code.
Have a good day! :)

Please open source code！

GAFF论文在FLIR数据集中每个类的AP实验结果

张博士你好，您在多光谱领域做出了很大的贡献！我目前刚刚入门这个领域，在做实验时候想要获取您的GAFF [1]、DAL [2]在FLIR数据集上的每个类别（Car，Person，Bicycle）的AP值，进行对比。
特此和您联系，十分希望能获得您当时论文中相应的实验结果。
mAP50 Car Person Bicycle
GAFF[1] 72.90% ？ ？ ？
DAL[2] 72.10% ？ ？ ？

[1] H. Zhang, E. Fromont, S. Lefevre, and B. Avignon, “Guided attentive feature fusion for multispectral pedestrian detection,” in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2021, pp. 72–80.
[2] H. Zhang, E. Fromont, S. Lefevre, and B. Avignon, “Deep active learning from multispectral data through cross-modality prediction inconsistency,” in 2021 IEEE International Conference on Image Processing (ICIP), 2021, pp. 449–453.

Problems of Speed and MR

Hi, I have some questions about your works, including CFR, GAFF and MD.

Firstly, according to your paper, GAFF runs 9.34 ms on 1080TI and 11.5ms on TX2, with feature fusion. In addition, in, your latest WACV 2022 paper, Low-Cost Multispectral Scene Analysis With Modality Distillation, speed of pure Resnet18+Retinanet with 640×512 input is 7ms.

However, I tried to reproduce Resnet18+Retinanet and VGG16+Retinanet, without feature fusion but with the same input size, the speeds are 18.3 ms and 35.6 ms, respectively. Note these speeds are tested on TITAN Xp, which is a GPU with even a little bit stronger power than 1080TI.

So my question is how did you test your models and report their speeds?

In addition, I think it's very hard to deploy a model from PC GPU such as 1080TI to edge device like TX2 with just a little loss of the speed, even if with FP16 or INT8 quantization.

Moreover, why the MR in your work MD inconsistent with your previous work GAFF, since the teacher network is exactly Resnet18+Retinanet+GAFF?

I think it would be helpful if you can release your entire code and model instead of results files. Thanks.

Open source code!

Hi, thank you for your great work! The work is so interesting, I hope the source code. Do you plan to open the source? If you don't have time to organize code, you can send the original code directly to my email. My e-mail is [[email protected]].

Open source code!

Hi, thank you for your great work! The work is so interesting, I hope the source code. Do you plan to open the source? If you don't have time to organize code, you can send the original code directly to my email. My e-mail is [email protected].