Image segmentation with touch interaction
In one embodiment, a method includes detecting one or more objects in an image, generating at least one mask for each of the detected objects, wherein each of the masks is defined by a perimeter, classifying the detected objects, receiving gesture input in relation to the image, determining whether one or more locations associated with the gesture input correlate with any of the masks, and providing feedback regarding the image in response to the gesture input. Each of the masks may include data identifying the corresponding detected object, and the perimeter of each mask may correspond to a perimeter of the corresponding detected object. The perimeter of the corresponding detected object may separate the detected object from one or more portions of the image that are distinct from the detected object.
1. A method comprising:
by a computing device, detecting one or more objects in an image;
by the computing device, generating at least one mask for each of the detected objects, wherein each of the masks is defined by a perimeter;
by the computing device, classifying the detected objects;
by the computing device, receiving gesture input in relation to the image;
by the computing device, determining whether one or more locations associated with the gesture input correlate with any of the masks; and
by a computing device, providing feedback regarding the image in response to the gesture input,
wherein a location associated with the gesture input corresponds to a region in which first and second masks overlap, and the feedback regarding the image comprises a description of a position of a first object that corresponds to the first mask relative to a second object that corresponds to the second mask.
2. The method of claim 1 , wherein each of the masks comprises data identifying the corresponding detected object, and the perimeter that defines each mask corresponds to a perimeter of the corresponding detected object.
3. The method of claim 2 , wherein the perimeter of the corresponding detected object separates the detected object from one or more portions of the image that are distinct from the detected object.
4. The method of claim 2 , wherein the mask comprises a matrix of data values that correspond to pixels of at least a portion of the image, pixels located inside the perimeter that defines the mask correspond to data values having a first value, and pixels located outside the perimeter that defines the mask correspond to data values having a second value.
5. The method of claim 1 , wherein the one or more objects correspond to one or more segments of the image, and the segments are detected by an image segmentation algorithm.
6. The method of claim 1 , wherein the one or more objects are classified as corresponding to a specified object type, and the feedback regarding the image comprises the specified object type.
7. The method of claim 6 , wherein the feedback is played as speech by an audio output component of the computing device.
8. The method of claim 6 , wherein the feedback is displayed as text by a display component of the computing device.
9. The method of claim 1 , wherein the one or more objects are classified as having a specified object name, and the feedback regarding the image comprises the specified object name.
10. The method of claim 1 , wherein the gesture input comprises a swipe gesture across a touch screen of the computing device or a tap gesture on the touch screen, and at least a portion of the gesture is detected at a location on the touch screen at which a portion of the image is displayed.
11. The method of claim 1 , wherein the gesture input comprises a pointing gesture that is detected as indicating a location on a display screen of the computing device at which a portion of the image is displayed.
12. The method of claim 1 , wherein determining whether one or more locations associated with the gesture input correlate with any of the masks comprises:
determining whether, for each of the detected objects, the one or more locations associated with the gesture input correspond to the mask for the detected object.
13. The method of claim 12 , wherein the one or more locations associated with the gesture input correspond to the mask for the detected object when the one or more locations are within the perimeter that defines the mask.
14. The method of claim 1 , wherein the first and second masks partially or fully overlap, the first and second objects are classified as having specified respective first and second object types, and
the feedback regarding the image further comprises the first and second object types.
15. The method of claim 1 , wherein the second object is smaller than the first object, the first and second objects are classified as having specified first and second object types, and
the feedback regarding the image further comprises an object type of the smaller of the first and second objects.
16. The method of claim 1 , wherein a location associated with the gesture input corresponds to a region in which the masks overlap, and providing feedback regarding the image in response to the gesture input comprises:
playing, by an audio component of the computing device, speech that includes an object type of the first object that corresponds to the first mask;
waiting for user input; and
in response to the user input, playing, by the audio component of the computing device, speech that includes an object type of the second object second object that corresponds to the second mask.
17. One or more computer-readable non-transitory storage media embodying software that is operable when executed to:
detect one or more objects in an image;
generate at least one mask for each of the detected objects, wherein each of the masks is defined by a perimeter;
classify the detected objects;
receive gesture input in relation to the image;
determine whether one or more locations associated with the gesture input correlate with any of the masks; and
provide feedback regarding the image in response to the gesture input,
wherein a location associated with the gesture input corresponds to a region in which first and second masks overlap, and the feedback regarding the image comprises a description of a position of a first object that corresponds to the first mask relative to a second object that corresponds to the second mask.
18. The media of claim 17 , wherein each of the masks comprises data identifying the corresponding detected object, and the perimeter that defines each mask corresponds to a perimeter of the corresponding detected object.
19. A system comprising: one or more processors; and a memory coupled to the processors comprising instructions executable by the processors, the processors being operable when executing the instructions to:
detect one or more objects in an image;
generate at least one mask for each of the detected objects, wherein each of the masks is defined by a perimeter;
classify the detected objects;
receive gesture input in relation to the image;
determine whether one or more locations associated with the gesture input correlate with any of the masks; and
provide feedback regarding the image in response to the gesture input,
wherein a location associated with the gesture input corresponds to a region in which first and second masks overlap, and the feedback regarding the image comprises a description of a position of a first object that corresponds to the first mask relative to a second object that corresponds to the second mask.
20. The system of claim 19 , wherein each of the masks comprises data identifying the corresponding detected object, and the perimeter that defines each mask corresponds to a perimeter of the corresponding detected object.