I have a couple thousand photos of whales taken from drones and I'm planning to build a simple binary classifier to run on these and future images to see if they contain a whale. I'd like to label specific pixels within the image as whale (1) or not whale (0) and I'm at a loss for a good free tool to do this. I'm imagining something where I can just zoom in and manually draw an outline around a whale but I'm open to other options. Does anyone know of a good way to do this?
Thanks!
If you are specifically looking to outline the whales, seems like FastAnnotationTool could work: https://github.com/christopher5106/FastAnnotationTool
Other options here: https://en.wikipedia.org/wiki/List_of_Manual_Image_Annotation_Tools
When had to annotate many images for a project, I built a fairly simple MATLAB gui that displayed images. I cycled through each image and if I clicked on a point on the image, that image was annotated and the point where I clicked was saved in a corresponding .mat file (.csv would have been better). It worked... mostly. I'd guess between python and OpenCV, this could be done easily without MATLAB.
Here's a great simple tool. It's fully in Python so you can play around with it to fit your needs more properly.
Give a shot to DataTurks It gives you simply slick UI + web based collaborative framework to work with your colleagues/mates to build the dataset.
Image from a sample Project which classifies images to celberity
Some more sample projects:
Opencv CVAT ( https://github.com/opencv/cvat) is a very new project with a long history. I've been playing around with it and its looks pretty good. It installs easily using docker-compose and has some growing pains, but its only been out for a few days.
