1. upload some image folders to server;
2. user can view the image and label at most 16 points describing what is the 3D outline for the truck in these pictures;
3. Above labeling process need some manually designating out points (for example, where is the upper and lower wheel is), and based on that create lines , parallel and vertical lines to create closed region.(There will be a picture to show that)
4. output the 3D outline points (8 point) coordinate data into XML file.
Please see more clarified requirement user story. Please do some evaluation based on this, please those who well understand the requirment and who has the right skill to finish it contact me and
do further communication.
1. upload one folder containing images to label to server direcotry
2. that folder will be used as a task index for those images waiting for labeling;
3. Every labeling data(xml file)will be stored in the same directory of image labelled;
4. that folder containing images and labelled output xml can be navigated and downloaded
5. every image status ( being labelled, not started to label, finished labeling) can be navigated, filtered, ordered, paginated, searched and whole status can be exported as excel file;
6. For labelling process:
a. provide tool box for : rectangle region labelling, single point labelling, line labelling.
a.1)rectangle region labelling toolbox: use it to select the rectangle region specified by top-left most and right-bottom most points. These two points should be outputed to final xml result file
a.2) single point labelling: only select one point with mouse click, the final coordinate in px units will be stored in the final xml result file
a.3) line labelling: use click two end points , then the line is created to span to outter most of the whole image region.
For those tool box, it should be easy to do edit job after relative entity created. For example, after single point labelling used, the single point can be moved, deleted by mouse activity or arrow key pressed. The rectangle labelling region can be changed by move one of the two points to cover different regions; For the line labelling toobox, lines can be changed by move or rotate or increase/decrease line length
b. provide convinient tool to create vertical line which passthrough one specific point or to create parallel line to one other line which passthrough one specific point
c. For those lines , there should be intersection for lines, the intersection point coordinate should be known by software, because some of these points will be the in the final xml file desigating the 3D region of the truck in the image.
d. After those manually adjusting lines, rectangles and drawn the 3D outline of the truck, we can produce final image plus only the 3D outline (lines connecting by 8 edge points), the 8 points coordinates should be outputed in the final xml file.
e. Special case for bigger truck in the image: If the intersection point of any line will not be in the image region, we should zoom in or zoom out to make sure the intersection point is in the original canvas(the original image size) and remember what is the zoom rate also.
37 freelancer đang chào giá trung bình ￥3800 cho công việc này
Hi! I have the most interest on your project. Your requirement really attract me to bid on your project. Its' my best skill so I think if u have good start with me everthing is easy. Thx for your attention.