Browser-based Image Gallery Annotation Interface (JavaScript based)

Improve existing JavaScript image annotation interface, and build experiment management backend that stores the interface configuration options and image URLs. Existing image annotation interface is available here: [login to view URL] . The PHP code is available on request. * Interface basic features (EXISTING) * - displays a gallery of images, using an input list of URLs (one per line) - ignores and does not display any incorrect URLs, or images that do not download - allows to select (by clicking on a thumbnail) images, and add them to a list of selected images - right click on an image opens it in its own tab - the gallery display settings can be customized: size of the thumbnails, number per line, number of URLs to read from the list To test it out, click submit while leaving the "Input image list (separated by line-breaks)" blank. This will load some sample image URLS. * User interface EXTENSIONS required (JavaScript/HTML) * - allow selecting multiple images at a time in gallery view e.g. starting from a selected image, user holds down shift and clicks on another image - option to display a fixed number of thumbnails per row (fixed row size, as done currently), or dynamically adjust the number depending on specified thumbnail size (width x height) - option to fit displayed images to thumbnail frame size, or center-crop to thumbnail size (current option) - fix the way long URLs are being displayed with the thumbnails; add configuration option to hide URLs The remaining JS functionality that is not updated should be kept as existing. * EXTENSIONS to backend * - allow reading configuration options from a local file in addition to the URLs - configuration parameters include the display settings, list of image URLs, which form an "experiment" - allow accessing an experiment (that corresponds to a local file name) via a direct link e.g. http://.../annotation_tool/experiment_name - after a user takes part in an experiment, and selects a number of images, she has the option to submit the selection (clicks a submit button); the server stores this selection, together with a unique user ID (or session ID); a user can also be required to enter a name at the time of submission, in order to identify herself (a configuration parameter makes this required or disabled) - the selected images from each participant in the experiment are saved to a database so that they can be easily retrieved later - configuration option to show (current behaviour) or hide the list of errors and selected URLs at the bottom of the gallery page - after submitting a task, a user is shown a unique code that he can use to identify his submission - if a user that has completed an experiment opens the same link again, he will not be allowed to retake it and his unique code will be shown - a configuration option enables/disables showing the unique code for each experiment Let's further discuss this to clarify. For instance, the choice of backend is flexible, it does not have to be PHP, Python works just as well. I would prefer to have this running as an hourly paid project. Thus, you should be comfortable with using the time-tracker. * MUST DO * Ensure you include the following in your bid message: 1. if you are comfortable using the time-tracker 2. a link or description of a previous project you worked on that is the most similar to the one proposed here I will not renegotiate the bid amount, please post your final bid from the start.