Stack Overflow
Stack Overflow

Questions tagged [faster-rcnn]

Faster-RCNN is a very common Regional Convolutional Neural Network architecture that detects and recognizes objects in images in a single forward pass

332 questions
14
votes
1 answer

What exactly are the losses in Matterport Mask-R-CNN?

I use Mask-R-CNN to train my data with it. When i use TensorBoard to see the result, i have the loss, mrcnn_bbox_loss, mrcnn_class_loss, mrcnn_mask_loss, rpn_bbox_loss, rpn_class_loss and all the same 6 loss for the validation: val_loss,…
Mob
  • 409
  • 1
  • 5
  • 14
8
votes
1 answer

Why does roi_align not seem to work in pytorch?

I am a pytorch beginner. It seems that there is a bug in the RoIAlign module in pytorch. The code is simple but the result is out of my expectation. code: import torch from torchvision.ops import RoIAlign if __name__ == '__main__': output_size…
sunshk1227
  • 219
  • 2
  • 12
7
votes
3 answers

TensorFlow 2 Mask-RCNN?

I can't seem to find a reliable version of Mask-RCNN for TensorFlow 2. The matterport mask-rcnn (https://github.com/matterport/Mask_RCNN) has depreciated Tensorflow 1 code. Does anyone know of TensorFlow 2 implementations of RCNN either based on…
myelin
  • 89
  • 1
  • 2
7
votes
1 answer

Does people train object detection methods with early stopping and what were their settings?

I'm working on some stuff related to object detection methods (YOLOv3, Faster-RCNN, RetinaNet, ... ) and I need to train on VOC2007 and VOC2012 (using pretrained models of course). However when I read the relevant papers I do not see people describe…
Dang Manh Truong
  • 795
  • 2
  • 10
  • 35
5
votes
2 answers

How to prune a Detectron2 model?

I'm a teacher who is studying computer vision for months. I was very excited when I was able to train my first object detection model using Detectron2's Faster R-CNN model. And it works like a charm! Super cool! But the problem is that, in order to…
knayman
  • 97
  • 8
5
votes
3 answers

Resnet-18 as backbone in Faster R-CNN

I code with pytorch and I want to use resnet-18 as backbone of Faster R-RCNN. When I print structure of resnet18, this is the output: >>import torch >>import torchvision >>import numpy as np >>import torchvision.models as models >>resnet18 =…
CVDE
  • 393
  • 3
  • 19
5
votes
1 answer

Implement Faster Rcnn from scratch

I want to build my own Faster-RCNN model from scratch for multi-object detection from image data. Can somebody please refer me good sources to step by step approach to implement faster-RCNN? Which one will be good YOLO or faster-RCNN in terms of…
gpu
  • 81
  • 1
  • 1
  • 4
4
votes
0 answers

Value Error: Expected x_max for bbox (0.73, 0.26, 1.02, 0.39),to be in the range [0.0, 1.0], got 1.029

When I train my own dataset with Faster RCNN by pytorch, I got this error - 'ValueError: Expected x_max for bbox (0.7343220683333334, 0.26440678624999997, 1.0292373053333335, 0.393220352, tensor(2)) to be in the range [0.0, 1.0], got…
Josekinkin
  • 41
  • 3
4
votes
1 answer

In training mode, targets should be passed

I am new to deep learning and have the project on detecting traffic lights in university where we can use open-source code. So, I tried to run the code on kaggle https://www.kaggle.com/endoruk1234/trafficlightdetection-fasterrcnn-pytorch/log However…
Angelina Parfenova
  • 43
  • 1
  • 3
4
votes
1 answer

High mAP@50 with low precision and recall. What does it mean and what metric should be more important?

I am comparing models for the detection of objects for maritime Search and Rescue (SAR) purposes. From the models that I used, I got the best results for the improved version of YOLOv3 for small object detection and for FASTER RCNN. For YOLOv3 I got…
Panicum
  • 794
  • 1
  • 9
  • 32
4
votes
2 answers

Adding multiple classes in Mask R-CNN

I am using Matterport Mask RCNN as my model and I'm trying to build my database for training. After much deliberation over the below problem, I think what I'm actually asking is how do I add more than one class (+ BG)? I get the following…
The Gibbinold
  • 259
  • 3
  • 12
4
votes
1 answer

Mask RCNN training on custom dataset hangs

I'm trying to train a Mask RCNN model on a custom dataset. The dataset I use for testing is the kangaroo dataset from https://github.com/experiencor/kangaroo and I'm essentially following this example:…
Yorian
  • 2,002
  • 5
  • 34
  • 60
4
votes
2 answers

Transfer learning in Pytorch using fasterrcnn_resnet50_fpn

I am looking for Object Detection for custom dataset in PyTorch. Tutorial here provides a snippet to use pre-trained model for custom object classification model_ft = models.resnet18(pretrained=True) num_ftrs = model_ft.fc.in_features model_ft.fc =…
addcolor
  • 455
  • 8
  • 23
4
votes
2 answers

py:445: UserWarning: Matplotlib is currently using agg, which is a non-GUI backend, so cannot show the figure. % get_backend())

I was trying to run the Mask-RCNN repository provided by the matterport in Github. https://github.com/matterport/Mask_RCNN. when I run the demo in the anaconda, it showed "C:\Anaconda\lib\site-packages\matplotlib\figure.py:445: UserWarning:…
JON
  • 41
  • 1
  • 2
3
votes
0 answers

pytorch faster r-cnn, no bounding box image return value error(all bounding boxes should have positive height and width)

I'm trying to train Faster R-CNN on custom dataset. When I train with no object image(image without bounding boxes), it returns value error 'All bounding boxes should have positive height and width'. My dataset contains image, target([xmin, ymin,…
Naeun Lee
  • 31
  • 2
1
2 3
…
22 23