Classification of three-dimensional (3D) objects in a virtual environment
Some implementations relate to methods, systems, and computer-readable media for classifying three-dimensional (3D) objects. In some implementations, the method includes generating a plurality of images of a candidate 3D object, determining one or more histogram of oriented gradients (HOG) vectors for each image, determining an asset feature of the candidate 3D object based on the one or more HOG vectors, determining if the asset feature of the candidate 3D object matches an authentic asset feature of at least one authentic 3D object, if the asset feature of the candidate 3D object matches the authentic asset feature of the at least one authentic 3D object, classifying the candidate 3D object as an inauthentic object, and if the asset feature of the candidate 3D object does not match the authentic asset feature of the at least one authentic 3D object, classifying the candidate 3D object as an authentic object.
1 . A computer-implemented method, comprising:
generating a plurality of images of a candidate 3D object, wherein each image of the plurality of images of the candidate 3D object is from a respective camera position of two or more camera positions;
determining one or more histogram of oriented gradients (HOG) vectors for each image of the plurality of images of the candidate 3D object;
determining an asset feature of the candidate 3D object based on the one or more HOG vectors for each of the plurality of images of the candidate 3D object;
determining if the asset feature of the candidate 3D object matches an authentic asset feature of at least one authentic 3D object by calculating a vector distance between the asset feature of the candidate 3D object and the authentic asset feature of the at least one authentic 3D object, wherein it is determined that the asset feature of the candidate 3D object matches the authentic asset feature of the at least one authentic 3D object if the vector distance meets a threshold vector distance;
if the asset feature of the candidate 3D object matches the authentic asset feature of the at least one authentic 3D object, classifying the candidate 3D object as an inauthentic object; and
if the asset feature of the candidate 3D object does not match the authentic asset feature of the at least one authentic 3D object, classifying the candidate 3D object as an authentic object.
2 . The computer-implemented method of claim 1 , wherein the vector distance is one or more of Euclidean distance, Manhattan distance, Hamming distance, cosine distance, or combinations thereof.
3 . The computer-implemented method of claim 1 , wherein determining if the asset feature of the candidate 3D object matches an authentic asset feature of at least one authentic 3D object further comprises performing a rotationally invariant comparison of the asset feature of the candidate 3D object and the authentic asset feature of the at least one authentic 3D object.
4 . The computer-implemented method of claim 3 , wherein performing the rotationally invariant comparison of the asset feature of the candidate 3D object and the authentic asset feature of the at least one authentic 3D object comprises:
generating a plurality of rolled asset feature vectors of the candidate 3D object based on the asset feature of the candidate 3D object, wherein each rolled asset feature vector corresponds to a particular orientation of the candidate 3D object; and
comparing each of the plurality of rolled asset feature vectors with the authentic asset feature of the at least one authentic 3D object.
5 . The computer-implemented method of claim 1 , wherein determining the one or more HOG vectors comprises determining one or more pyramidal HOG vectors, and wherein each of the one or more pyramidal HOG vectors is generated by concatenating at least two HOG vectors of the candidate 3D object generated at multiple resolutions of a respective image of the plurality of images of the candidate 3D object.
6 . The computer-implemented method of claim 1 , wherein generating the plurality of images of the candidate 3D object comprises generating the plurality of images at one or more azimuth and elevation points.
7 . The computer-implemented method of claim 1 , wherein generating at least one image of the plurality of images of the candidate 3D object comprises adjusting a camera view during capture of the at least one image such that the candidate 3D object occupies at least a specified area of the image.
8 . The computer-implemented method of claim 1 , further comprising prior to generating the plurality of images of the candidate 3D object, replacing a texture of the candidate 3D object with a white plastic material.
9 . The computer-implemented method of claim 1 , wherein classifying the candidate 3D object as the authentic object further comprises assigning a flag to the candidate 3D object, and wherein the flag is readable by a game engine and causes the game engine to enable use of the candidate 3D object in a virtual environment hosted by the game engine.
10 . The computer-implemented method of claim 1 , wherein classifying the candidate 3D object as the inauthentic object further comprises assigning a flag to the candidate 3D object, and wherein the flag is readable by a game engine and causes the game engine to prevent use of the candidate 3D object in a virtual environment hosted by the game engine.
11 . The computer-implemented method of claim 1 , further comprising performing principal component analysis of the asset feature to reduce a dimension of the asset feature, and wherein determining if the asset feature of the candidate 3D object matches the authentic asset feature of the at least one authentic 3D object comprises performing a comparison of a reduced dimension asset feature of the candidate 3D object against a reduced dimension authentic asset feature of the at least one authentic 3D object.
12 . A non-transitory computer-readable medium comprising instructions that, responsive to execution by a processing device, cause the processing device to perform operations comprising:
generating a plurality of images of a candidate 3D object, wherein each image of the plurality of images of the candidate 3D object is from a respective camera position of two or more camera positions;
determining one or more histogram of oriented gradients (HOG) vectors for each image of the plurality of images of the candidate 3D object;
determining an asset feature of the candidate 3D object based on the one or more HOG vectors for each of the plurality of images of the candidate 3D object;
determining if the asset feature of the candidate 3D object matches an authentic asset feature of at least one authentic 3D object by calculating a vector distance between the asset feature of the candidate 3D object and the authentic asset feature of the at least one authentic 3D object, wherein it is determined that the asset feature of the candidate 3D object matches the authentic asset feature of the at least one authentic 3D object if the vector distance meets a threshold vector distance;
if the asset feature of the candidate 3D object matches the authentic asset feature of the at least one authentic 3D object, classifying the candidate 3D object as an inauthentic object; and
if the asset feature of the candidate 3D object does not match the authentic asset feature of the at least one authentic 3D object, classifying the candidate 3D object as an authentic object.
13 . The non-transitory computer-readable medium of claim 12 , wherein the vector distance is one or more of Euclidean distance, Manhattan distance, Hamming distance, cosine distance, or combinations thereof.
14 . The non-transitory computer-readable medium of claim 12 , wherein determining the one or more HOG vectors comprises determining one or more pyramidal HOG vectors, and wherein each of the one or more pyramidal HOG vectors is generated by concatenating at least two HOG vectors of the candidate 3D object generated at multiple resolutions of a respective image of the plurality of images of the candidate 3D object.
15 . A system comprising:
a memory with instructions stored thereon; and
a processing device, coupled to the memory, the processing device configured to access the memory and execute the instructions, wherein the instructions cause the processing device to perform operations including:
generating a plurality of images of a candidate 3D object, wherein each image of the plurality of images of the candidate 3D object is from a respective camera position of two or more camera positions;
determining one or more histogram of oriented gradients (HOG) vectors for each image of the plurality of images of the candidate 3D object;
determining an asset feature of the candidate 3D object based on the one or more HOG vectors for each of the plurality of images of the candidate 3D object;
determining if the asset feature of the candidate 3D object matches an authentic asset feature of at least one authentic 3D object by calculating a vector distance between the asset feature of the candidate 3D object and the authentic asset feature of the at least one authentic 3D object, wherein it is determined that the asset feature of the candidate 3D object matches the authentic asset feature of the at least one authentic 3D object if the vector distance meets a threshold vector distance;
if the asset feature of the candidate 3D object matches the authentic asset feature of the at least one authentic 3D object, classifying the candidate 3D object as an inauthentic object; and
if the asset feature of the candidate 3D object does not match the authentic asset feature of the at least one authentic 3D object, classifying the candidate 3D object as an authentic object.
16 . The system of claim 15 , wherein the vector distance is one or more of Euclidean distance, Manhattan distance, Hamming distance, cosine distance, or combinations thereof.
17 . The system of claim 15 , wherein determining the one or more HOG vectors comprises determining one or more pyramidal HOG vectors, and wherein each of the one or more pyramidal HOG vectors is generated by concatenating at least two HOG vectors of the candidate 3D object generated at multiple resolutions of a respective image of the plurality of images of the candidate 3D object.
18 . The system of claim 15 , wherein determining if the asset feature of the candidate 3D object matches an authentic asset feature of at least one authentic 3D object further comprises performing a rotationally invariant comparison of the asset feature of the candidate 3D object and the authentic asset feature of the at least one authentic 3D object.
19 . The system of claim 15 , wherein the operations further comprise, prior to generating the plurality of images of the candidate 3D object, replacing a texture of the candidate 3D object with a white plastic material.
20 . The system of claim 15 , wherein the operations further comprise reducing a dimension of the asset feature, and wherein determining if the asset feature of the candidate 3D object matches the authentic asset feature of the at least one authentic 3D object comprises performing a comparison of a reduced dimension asset feature of the candidate 3D object against a reduced dimension authentic asset feature of the at least one authentic 3D object.