web-based image annotation
Allows sub-areas of an image to be manually defined. This is useful for observational-based psychology studies or creating training data for object/scene detection algorithms.
It uses a minimal set of tools: HTML5, PHP, and javascript. As such, it can easily be deployed on any basic web-server.
Users can click squares defined by the grid-overlay to highlight portions of the image. Output includes a list of indices, indicating the set of squares clicked.
Here is an example output file
#config
pixel_size,40
image_width,1280
image_height,720
highlight_color,0,255,255
highlight_transparency,0.5
randomize_images,0
line_width,1
radio_buttons,0
textbox_label,notes if any
results_save,1
savefilename,../results/demo.txt
confirmation_code,demo
taskcode,demo
#selection
../images/demo/img_127.jpg,208,209,223,224
../images/demo/img_215.jpg,193,211,225,226,229
../images/demo/img_3.jpg,137,154,155,172,173
../images/demo/img_355.jpg,283,284,285
../images/demo/img_537.jpg,185,186,187,203,204
#end
#textbox
../images/demo/img_127.jpg,horse
../images/demo/img_215.jpg,dog
../images/demo/img_3.jpg,cat
../images/demo/img_355.jpg,barn
../images/demo/img_537.jpg,apple
#end