Method for generating labeling data that describe an image content of images depicting at least one scene, corresponding processing device, vehicle and data storage medium
A method for generating labeling data is disclosed that describes an image content of images depicting at least one scene, wherein in a processing device image data are received from an imaging and a segmentation unit that detects at least one object in the image data. A graphical processing unit generates a respective graphical object marker that marks the at least one detected object and a display control unit displays an overlay of the at least one scene and the at least one object marker. An input reception unit receives a respective user input for each object marker, wherein the respective user input provides the image content of the image region marked by the object marker.
1. A method for generating labeling data that describe an image content of images depicting at least one scene, wherein in a processing device:
image data describing the images are received from an imaging unit;
a segmentation unit detects at least one object in the at least one scene on the basis of the image data and generates segmentation data describing a respective image region that depicts the respective detected object in the images;
a graphical processing unit generates marker data that describe a respective graphical object marker that marks the respective image region of the at least one detected object on the basis of the segmentation data;
a display control unit displays an overlay of the at least one scene and the at least one object marker on the basis of the marker data, wherein each object marker is respectively positioned over the respective object that it marks and wherein the overlay is displayed by controlling at least one display unit;
an input reception unit receives a respective user input for each object marker, wherein the respective user input provides a description of the image content of the image region marked by the object marker; and
a label generation unit generates the labeling data, wherein by the labeling data the respective image region depicting a respective object is associated with the description of the object as provided by the user input.
2. The method according to claim 1 , wherein at least one of the following display units is controlled:
a head-up unit, wherein the head-up unit displays the at least one object marker on a transparent combiner screen,
a monitor screen that displays both the image data and the marker data, and/or
a set of augmented reality glasses.
3. The method according to claim 1 , wherein the input reception unit receives the user input as a voice command and performs a speech recognition for recognizing the image content and/or wherein the input reception unit receives the user input as a typed input and/or wherein the input reception unit displays a list of possible object descriptors and received a user selection of one of the object descriptors as the user input.
4. The method according to claim 1 , wherein the input reception unit receives a gesture signal from a gesture recognition device and/or a gaze direction signal from a gaze recognition device and detects a marker selection that indicates which object marker the user input refers to on the basis of the gesture signal and/or the gaze direction signal.
5. The method according to claim 1 , wherein in the case of several detected objects the corresponding object markers are displayed sequentially.
6. The method according to claim 1 , wherein at least one additional user input is received that provides segmentation data of a missed object that was missed by the segmentation unit.
7. The method according to claim 1 , wherein the at least one object is detected during a test drive of a driving vehicle and the object marker is displayed in the vehicle and a user providing the user input is situated in the vehicle during the test drive.
8. The method according to claim 7 , wherein the vehicle is an autonomously driving vehicle and the path the vehicle plans to take and/or the abstracted map data containing environment information is displayed.
9. The method according to claim 1 , wherein on the basis of the image data and the labeling data an artificial neural network is trained to recognize the objects in the image data.
10. A processing device comprising:
at least one processor; and
a data storage medium coupled to the at least one processor,
wherein the data storage medium stores computer readable instructions that cause the at least one processor to perform a method according to claim 1 if executed by the at least one processor.
11. An autonomously driving vehicle comprising the processing device according to claim 10 .
12. A non-transitory computer readable data storage medium, wherein the data storage medium stores computer readable instructions that cause at least one processor to perform a method according to claim 1 , if executed by the at least one processor.