AlignPS icon indicating copy to clipboard operation
AlignPS copied to clipboard

Code for CVPR 2021 paper: Anchor-Free Person Search

Results 10 AlignPS issues
Sort by recently updated
recently updated
newest added

1.我理解的对吗? 1.Am I right to understand that? 假如同一个人,他的大图和小图分别对应不同层次的特征图。不同的层次特征图肯定不完全一样。同一个人不同大小的图片的特征竟然在不同的特征图上不一样,很可能认为为不是一个人。**如果query是个小图,是高层次的特征,gallery是大图,使用低层次的特,由于两个特征来自不同尺度的特征图,因此有可能导致特征向量点乘不正常,从而造成误判。所以你只用了一层特征图。** If it is the same person, his large picture and small picture correspond to different levels of feature maps. Different hierarchical feature...

Hi: I encountered some problems when running the source code: 2022-02-17 14:43:37,319 - mmdet - INFO - Epoch [1][50/2852] lr: 3.987e-04, eta: 2 days, 6:05:49, time: 1.423, data_time: 0.068, memory:...

writing results to work_dirs/fcos_center-normbbox-centeronreg-giou_r50_caffe_fpn_gn-head_dcn_4x4_1x_cuhk_reid_1500_stage1_fpncat_dcn_epoch24_multiscale_focal_x4_bg-2_lconv3dcn_sub_triqueue_dcn0/results_1000.pkl ------------------------ threshold: 0.2 mAP = 0.00% Top- 1 = 0.00% Top- 5 = 0.00% Top-10 = 0.00% fcos_center-normbbox-centeronreg-giou_r50_caffe_fpn_gn-head_dcn_4x4_1x_cuhk_reid_1500_stage1_fpncat_dcn_epoch24_multiscale_focal_x4_bg-2_lconv3dcn_sub_triqueue_dcn0

@erjanmx @daodaofr hi thanks for open sourcing the code base, i have a few queries on running the inference 1. I have downloaded the dataset of Chuk-sys and i have...

Hello, I wonder how to visualize the offset of deformable conv? Could you offer a demo code?

hello,thanks for your project. i want to know whether to support distributed training. And what should i do to make it support distributed training.

Does this workspace support the implementation of the journal version paper? [Efficient Person Search: An Anchor-Free Approach](https://arxiv.org/abs/2109.00211).

here I only find how to change the gallery_size, but I don't know how to use the Occlusion and Resolution on CUHK dataset, Have you done research in this area,...

i am currently testing to build the image using docker in a virtual machine, if so does it require any gpu usage?

Hi, I found a directory named `convert_datasets` where I further found a script to convert CUHK dataset to COCO format, so I wonder, whether there is a script to convert...