I have data labeled with 3D bounding boxes (KITTI object detection 3D), and I want to apply one of the Tensorflow object detection models (such as Faster R-CNN) to perform object detection. So far, All of the examples I saw are using 2D Bounding boxes. Does this mean that a 3D box is not allowed by such API?
If NOT, does Anyone has an example using a 3D bounding box example model?
Related
I have a set of icons and a screen recording, the icons are not annotated and have no bounding boxes, they are just png icons with image-level labels eg: "instagram", "facebook", "chrome".
The task is to find the icons within the screen recording and draw a bounding box around them, given the above prerequistes.
My idea of approach so far is:
Use selective search to find ROIs
Use a CNN to classify the regions
Filter out non-icons regions
Draw bounding boxes around positive labelled ROIs
Use resulting screen images with bounding boxes to train a FAST R-CNN
but I am stuck at step 2, I have no idea on how to train the CNN with the image-level labelled icons.
If I make a dataset of all the possible icon-images, with no background or context informations, is it possible to train the CNN to answer the question "Does the ROI includes a possible icon?"
Is it possible to get the results or coordinates of the mask detection or the bounding box surrounding the image? I am using Mask R-CNN from matterport and the visualization of the masks on the image works quite good, but I woukd like to save the coordinates.
I am not sure how you are using this model. But when you import their model and use the detect method (which is the straightforward way to use their model), then the coordinates are returned immediately.
See for an explanation of what is returned from model.detect this documentation.
In short, per image you get a dict and your coordinates will be the_dict["rois"][1].
My problem is not exactly annotate data using polygon, circle or line, it's how to use these annotated data to gerenate a ".tfrecord" file and perform an object detection. The tutorials I saw use rectangle annotation, like these: taylor swift detection raccon detection
It would be a great one for me if the objects I want to detect (pipelines) were not too close.
Example of rectangle drawn in PASCAL VOC format:
<bndbox>
<xmin>82</xmin>
<xmax>172</xmax>
<ymin>108</ymin>
<ymax>146</ymax>
</bndbox>
Is there a way to add a "mask" to highlight some part of this bounding box?
If it's something unclear, please let me know.
You can go for instance segmentation instead of object detection if your objects are very close to each other, there you can use polygons to generate masks and bounding boxes to train the model.
Consider this well presented and easy to use repository for mask-rcnn(kind of instance segmentation)
https://github.com/matterport/Mask_RCNN
check this for lite weight mask-rcnn
I am doing an object detection on sample toy sets using tensorflow RCNN on my own dataset.
Sometimes when two toys are closed to each other, the tensor flow is bounding both boxes together. So instead of returning two bounding boxes, tensorflow returning one bigger bounding box.
(as in the attached image both red and blue cars have been bounded together)
Does anyone know what is the issue here and how to do individual identification?
Detected Output image
Tensorflow's Object Detection API has an option in the .config file to add an keep_aspect_ratio_resizer. If I resize my training data using this, will the corresponding bounding boxes be resized as well? If they don't match up then the network is seeing incorrect examples.
Yes, the boxes will be resized to be compatible with the images as well!