search

tianzhi0549 / FCOS

FCOS: Fully Convolutional One-Stage Object Detection (ICCV'19)
https://arxiv.org/abs/1904.01355
Other
3.27k stars 630 forks source link

Convertion of segmentation to fcos #367

Open abhigoku10 opened 3 years ago

abhigoku10 commented 3 years ago

@tianzhi0549 thanks for opensourcing the code i am trying to convert any anchor based segmentation architecture to anchor free architecture i was able to perform the conversion but getting bad results so i double checked the debugged difference and below is the observation

  1. the input image size in fcos is minimum 800 but in yolact its set to 544
  2. THe feature map wxh of fcos starts with 136x104 and ends with 9x7 but with yolact its 68x68 to 5x5
  3. The strides which are used in fcos are strides = [8, 16, 32, 64, 128] which is similar to present in yolact "fpn_fm_shape = [math.ceil(cfg.img_size / stride) for stride in (8, 16, 32, 64, 128)]"
  4. In Fcos only flip and padding augmentations are done but in yolact many augmentations are done but i commented the t_0_1box () ie normalization of the bounding box

Can you share your thoughts on these points