Add Pascal VOC Datasets for image detection and segmentation #209
Conversation
… years for seg dataset mlverse#139
|
@cregouby the original website is down. |
pascal_seg <- pascal_segmentation_dataset(transform = transform_to_tensor, download = TRUE)
# Access the first image and its mask
first_item <- pascal_seg[1]
first_item$x # Image
first_item$y$masks # Segmentation mask
# Visualise the first image and its mask
masked_img <- draw_segmentation_masks(first_item)
tensor_image_browse(masked_img)
pascal_det <- pascal_detection_dataset(transform = transform_to_tensor, download = TRUE)
# Access the first image and its bounding boxes
first_item <- pascal_det[1]
first_item$x # Image
first_item$y$labels # Object labels
first_item$y$boxes # Bounding box tensor
# Visualise the first image with bounding boxes
boxed_img <- draw_bounding_boxes(first_item)
tensor_image_browse(boxed_img)
|
|
@cregouby This PR is ready for review. |
cregouby
left a comment
cregouby
left a comment
There was a problem hiding this comment.
praise Nice shoot, with very effective example. Thanks !
todo see inline
Co-authored-by: cregouby <cregouby@users.noreply.github.com>
…shed/torchvision into feat/pascalDatasets
cregouby
left a comment
cregouby
left a comment
There was a problem hiding this comment.
todo major performance improvement is needed. Some hints are included.
cregouby
left a comment
cregouby
left a comment
There was a problem hiding this comment.
praise This is much better now. Thanks for the update
todo see inline
| classes = c( | ||
| "background", "aeroplane", "bicycle", "bird", "boat", "bottle", | ||
| "bus", "car", "cat", "chair", "cow", "diningtable", "dog", | ||
| "horse", "motorbike", "person", "pottedplant", "sheep", | ||
| "sofa", "train", "tvmonitor" | ||
| ), |
There was a problem hiding this comment.
todo could you switch to classes=voc_segmentation_classes as currently there are inconsistencies with yours :
There was a problem hiding this comment.
classes = c(
"background", "aeroplane", "bicycle", "bird", "boat", "bottle",
"bus", "car", "cat", "chair", "cow", "dining table", "dog",
"horse", "motorbike", "person", "potted plant", "sheep",
"sofa", "train", "tv"
),There was a problem hiding this comment.
I have made the above change. It would help if I knew where the voc_segmentation_classes are from so I can make them consistent.
2 participants
This PR adds two datasets:
pascal_segmentation_dataset()andpascal_detection_dataset()Closes #139