DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Claim Rejections - 35 USC § 102
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claims 1, 4-7, 9, 11, 14, 15, and 18 are rejected under 35 U.S.C. 102(a)(2) as being anticipated by Jain et al. (US Pub. 2025/0095221), hereinafter Jain.
Regarding claim 1, Jain discloses a method for wallpaper setting, applicable to an electronic device having a display and comprising: obtaining at least two wallpaper setting conditions (Paragraph [0003]: image tags, the textual information, and/or the extracted features are provided as input to a generative text model, which is trained using machine learning to transform unformatted text into text prompts that inform a generative image model how to generate a background image for the object. Accordingly, the generative text model generates a prompt and provides the prompt as input to a generative image model, which is trained using machine learning to generate background images based on the prompt. Upon receiving the prompt and an image of the multiple images, the generative image model generates a background image based on the prompt. Further, an output image is generated by incorporating the object (as depicted in one of the multiple images) into the background image; Paragraph [0056]: Using the entity extraction technique, the feature extraction module 302 outputs features 310 including one or more of the entities 306 that match contiguous portions of text in the input data 114 and/or one or more additional entities 306 to which the matching entities 306 are connected in the knowledge graph 304. In an example, the feature extraction module 302 identifies the term “umbrella” that is present as an entity 306 in the knowledge graph 304, and in the input data 114. Further, the entity 306 “umbrella” is connected via an association 308 to the entity 306 “patio table.” In this example, the feature extraction module 302 outputs, as the features 310, the entity 306 “umbrella” and the entity 306 “patio table” despite the phrase “patio table” not occurring in the input data 114); comparing the at least two wallpaper setting conditions with information tags of wallpapers in a wallpaper database, and matching out, from the wallpaper database, candidate wallpapers each of which has information tags comprising the at least two wallpaper setting conditions, wherein the candidate wallpapers are generated through a preset wallpaper generation model according to a wallpaper generation sentence associated with the at least two wallpaper setting conditions (Fig. 3; Paragraph [0060]: a generative text model 316 receives, as input, the features 310, the predicted usage 314, and/or the unprocessed input data 114. In particular, the unprocessed input data 114 received by the generative text model 316 includes the image tags 234, the textual information 120, and/or the user data 210. Broadly, the generative text model 316 is a machine learning text-to-text model that is trained to take unformatted text as input, and output a prompt 318 that is suitable for background image generation by a generative image model 320. Notably, the generative text model 316 and the generative image model 320 collectively form the generative AI models; Paragraphs [0065]-[0066]: generative text model 316 is trained by comparing the generated prompt 318 to the corresponding target prompt. By way of example, the generated prompt 318 and the target prompt are converted to vectors using a text vectorization technique (e.g., Term Frequency-Inverse Document Frequency (TF-IDF), Word2Vec, and GloVe), and compared using a similarity metric (e.g., cosine similarity or Euclidean distance). Parameters of the generative text model 316 are iteratively updated based on degrees of similarity between the generated prompt 318 and the target prompt on different pairs of training data…the generative image model 320 receives the prompt 318 and the input data 114 as input. In particular, the input data 114 provided to the generative image model 320 is an image 116 depicting the object 118 for which the background image 126 is to be generated. Broadly, the generative image model 320 is a machine learning text-to-image model that is trained to receive a prompt 318 and an image 116 as input, and generate an output image 124 having a new background that is based on the prompt 318. To do so, the generative image model 320 generates a background image 126 based on the prompt 318. Further, the generative image model 320 replaces a background (as depicted in the image 116) with the background image 126, thereby incorporating the object (as depicted in the image 116) into the background image); and displaying at least part of the candidate wallpapers on the display of the electronic device (Fig. 3; Paragraph [0076]: After the visual saliency 130 is determined, the visual saliency module 128 compares the visual saliency 130 to a visual saliency threshold 132. By way of example, the visual saliency threshold 132 defines a minimum value for the visual saliency 130 that is to be achieved in order for the output image 124 to be output by the background generation system 112. Thus, if the visual saliency 130 satisfies the visual saliency threshold 132, the output image 124 is output, e.g., by integrating the output image 124 into a publication 212 for display in a user interface of the client device 204 requesting to access the publication 212. If the visual saliency 130 does not satisfy the visual saliency threshold 132, however, the visual saliency module 128 instructs the generative image model 320 to generate a different output image 124).
Regarding claim 4, Jain discloses the method for wallpaper setting of claim 1, wherein the candidate wallpapers are implemented as a plurality of candidate wallpapers, and displaying the at least part of the candidate wallpapers on the display of the electronic device comprises: displaying all the candidate wallpapers on the display of the electronic device in a random order (Paragraph [0086]: the example user interface 402 of FIG. 4b is displayed in response to a user input selecting the output image 124 displayed in the example user interface 400 of FIG. 4a. As shown in FIG. 4b, the example user interface 402 includes a first panel 410, a second panel 412, and a third panel 414. Further, the first panel 410 displays the images 116 currently being used in association with the publication 212. Additionally, the first panel 410 displays the output images 124 automatically generated by the background generation system 112 as candidates for replacing or supplementing the images 116 for use in association with the publication 212. As shown in the illustrated example, each respective image (e.g., including the images 116 and the output images 124) includes a visual indication of a utility score for the respective image as generated by the utility scoring model 404. In response to a user input selecting one of the output images 124, the selected output image 124 is rendered for display in the larger display area of the second panel 412).
Regarding claim 5, Jain discloses the method for wallpaper setting of claim 1, wherein after displaying the at least part of the candidate wallpapers on the display of the electronic device, the method for wallpaper setting further comprises: in response to a wallpaper switching instruction being still received after each of the candidate wallpapers is displayed, performing at least semantic association on the at least two wallpaper setting conditions, and directly invoking the preset wallpaper generation model to generate, according to a semantic association result, a new wallpaper (Paragraphs [0065]-[0066]: generative text model 316 is trained by comparing the generated prompt 318 to the corresponding target prompt. By way of example, the generated prompt 318 and the target prompt are converted to vectors using a text vectorization technique (e.g., Term Frequency-Inverse Document Frequency (TF-IDF), Word2Vec, and GloVe), and compared using a similarity metric (e.g., cosine similarity or Euclidean distance). Parameters of the generative text model 316 are iteratively updated based on degrees of similarity between the generated prompt 318 and the target prompt on different pairs of training data…the generative image model 320 receives the prompt 318 and the input data 114 as input. In particular, the input data 114 provided to the generative image model 320 is an image 116 depicting the object 118 for which the background image 126 is to be generated. Broadly, the generative image model 320 is a machine learning text-to-image model that is trained to receive a prompt 318 and an image 116 as input, and generate an output image 124 having a new background that is based on the prompt 318. To do so, the generative image model 320 generates a background image 126 based on the prompt 318. Further, the generative image model 320 replaces a background (as depicted in the image 116) with the background image 126, thereby incorporating the object (as depicted in the image 116) into the background image).
Regarding claim 6, Jain discloses the method for wallpaper setting of claim 1, wherein after displaying the at least part of the candidate wallpapers on the display of the electronic device, the method for wallpaper setting further comprises: in response to a wallpaper switching instruction being still received after each of the candidate wallpapers is displayed, determining whether computing power of the electronic device is less than a preset computing power threshold; in response to the computing power of the electronic device being greater than or equal to the preset computing power threshold, performing at least semantic association on the at least two wallpaper setting conditions, and directly invoking the preset wallpaper generation model to generate, according to a semantic association result, a new wallpaper (Paragraphs [0065]-[0066]: generative text model 316 is trained by comparing the generated prompt 318 to the corresponding target prompt. By way of example, the generated prompt 318 and the target prompt are converted to vectors using a text vectorization technique (e.g., Term Frequency-Inverse Document Frequency (TF-IDF), Word2Vec, and GloVe), and compared using a similarity metric (e.g., cosine similarity or Euclidean distance). Parameters of the generative text model 316 are iteratively updated based on degrees of similarity between the generated prompt 318 and the target prompt on different pairs of training data…the generative image model 320 receives the prompt 318 and the input data 114 as input. In particular, the input data 114 provided to the generative image model 320 is an image 116 depicting the object 118 for which the background image 126 is to be generated. Broadly, the generative image model 320 is a machine learning text-to-image model that is trained to receive a prompt 318 and an image 116 as input, and generate an output image 124 having a new background that is based on the prompt 318. To do so, the generative image model 320 generates a background image 126 based on the prompt 318. Further, the generative image model 320 replaces a background (as depicted in the image 116) with the background image 126, thereby incorporating the object (as depicted in the image 116) into the background image; Paragraph [0119]: cloud 714 includes and/or is representative of a platform 716 for resources 718. The platform 716 abstracts underlying functionality of hardware (e.g., servers) and software resources of the cloud 714. The resources 718 include applications and/or data that can be utilized while computer processing is executed on servers that are remote from the computing device 702. Resources 718 can also include services provided over the Internet and/or through a subscriber network, such as a cellular or Wi-Fi network); and in response to the computing power of the electronic device being less than the preset computing power threshold, sending the at least two wallpaper setting conditions to a cloud server, to make the cloud server perform at least semantic association on the at least two wallpaper setting conditions and invoke the preset wallpaper generation model to generate, according to the semantic association result, the new wallpaper; and obtaining the new wallpaper generated by the cloud server (Fig. 7; Paragraphs [0034]-[0036]: FIG. 1 is an illustration of an environment 100 in an example implementation that is operable to employ techniques described herein for saliency-based background generation. The illustrated environment 100 includes a computing device 102 and a client device 104 that are communicatively coupled, one to another, via a network 106. The computing device 102 and/or the client device 104 are configurable in a variety of ways. For instance, the computing device 102 and/or the client device 104 are configurable as a desktop computer, a laptop computer, a mobile device (e.g., assuming a handheld configuration such as a tablet or mobile phone as illustrated), and so forth. Thus, the computing device 102 and/or the client device 104 range from full resource devices with substantial memory and processor resources (e.g., personal computers, game consoles) to a low-resource device with limited memory and/or processing resources (e.g., mobile devices). Additionally, although a single computing device 102 is shown, the computing device 102 is also representative of a plurality of different devices, such as multiple servers utilized by a business to perform operations “over the cloud” as described in FIG. 7… computing device 102 is illustrated as including an image processing system 108. The image processing system 108 is implemented at least partially in hardware of the computing device 102 to process and transform digital images. Such processing includes creation of the digital images, modification of the digital images, and rendering of the digital images in a user interface 110 for output, e.g., by a display of the client device 104. Although illustrated as implemented locally at the computing device 102, functionality of the image processing system 108 is also configurable as whole or part via functionality available via the network 106, such as part of a web service or “in the cloud.”…example of functionality incorporated by the image processing system 108 to process the digital images is illustrated as a background generation system 112. As shown, the background generation system 112 receives, as input data 114, multiple images 116 depicting an object 118 and textual information 120 describing the object 118. In one or more implementations, the input data 114 is obtained from an online publication of the object 118 made available via an online publication service. In the illustrated example, for instance, a listing (e.g., a publication) of an umbrella (e.g., the object 118) is offered for sale via an online marketplace (e.g., an online publication service). Further, the listing includes multiple images 116 depicting the umbrella as well as textual information 120 describing the umbrella, e.g., a title of the umbrella, a description of the umbrella, and reviews of the umbrella).
Regarding claim 7, Jain discloses the method for wallpaper setting of claim 5, wherein performing at least semantic association on the at least two wallpaper setting conditions and directly invoking the preset wallpaper generation model to generate, according to the semantic association result, the new wallpaper comprise: identifying a style of icons on a display interface currently displayed on the display of the electronic device, to obtain corresponding icon-style information (Paragraph [0089]: he third panel 414 includes a user interface element 420 that is manipulable to specify an aspect ratio for the output images 124, and a user interface element 422 that is manipulable to specify one or more styles for the output images 124. The styles that are specifiable by the user interface element 422 include, for example, a degree of realism for the output images 124 (e.g., photorealistic images or cartoon images), color schemes for the output images 124, and the like. The third panel 414 additionally includes a user interface element 424 that is selectable to instruct the background generation system 112 to generate a new output image 124 in accordance with the updated parameters specified via the third panel 414. In response to receiving user input selecting the user interface element 424, for instance, the client device 204 communicates the updated parameters to the background generation system 112, which generates a new output image 124 in accordance with the techniques discussed herein based on the updated parameters. The new output image 124 is communicated back to the client device 204, which renders the new output image 124 for display in the first panel 410 and/or the second panel 412); and performing semantic association on the icon-style information and the at least two wallpaper setting conditions, and generating, through the preset wallpaper generation model, a candidate wallpaper according to the semantic association result (Paragraph [0089]: he third panel 414 includes a user interface element 420 that is manipulable to specify an aspect ratio for the output images 124, and a user interface element 422 that is manipulable to specify one or more styles for the output images 124. The styles that are specifiable by the user interface element 422 include, for example, a degree of realism for the output images 124 (e.g., photorealistic images or cartoon images), color schemes for the output images 124, and the like. The third panel 414 additionally includes a user interface element 424 that is selectable to instruct the background generation system 112 to generate a new output image 124 in accordance with the updated parameters specified via the third panel 414. In response to receiving user input selecting the user interface element 424, for instance, the client device 204 communicates the updated parameters to the background generation system 112, which generates a new output image 124 in accordance with the techniques discussed herein based on the updated parameters. The new output image 124 is communicated back to the client device 204, which renders the new output image 124 for display in the first panel 410 and/or the second panel 412).
Regarding claim 9, Jain discloses the method for wallpaper setting of claim 5, wherein performing at least semantic association on the at least two wallpaper setting conditions and directly invoking the preset wallpaper generation model to generate, according to the semantic association result, the new wallpaper comprise: performing at least semantic association on the at least two wallpaper setting conditions, and generating, through the preset wallpaper generation mode, a candidate wallpaper according to the semantic association result (Paragraph [0003]: image tags, the textual information, and/or the extracted features are provided as input to a generative text model, which is trained using machine learning to transform unformatted text into text prompts that inform a generative image model how to generate a background image for the object. Accordingly, the generative text model generates a prompt and provides the prompt as input to a generative image model, which is trained using machine learning to generate background images based on the prompt. Upon receiving the prompt and an image of the multiple images, the generative image model generates a background image based on the prompt. Further, an output image is generated by incorporating the object (as depicted in one of the multiple images) into the background image; Paragraph [0056]: Using the entity extraction technique, the feature extraction module 302 outputs features 310 including one or more of the entities 306 that match contiguous portions of text in the input data 114 and/or one or more additional entities 306 to which the matching entities 306 are connected in the knowledge graph 304. In an example, the feature extraction module 302 identifies the term “umbrella” that is present as an entity 306 in the knowledge graph 304, and in the input data 114. Further, the entity 306 “umbrella” is connected via an association 308 to the entity 306 “patio table.” In this example, the feature extraction module 302 outputs, as the features 310, the entity 306 “umbrella” and the entity 306 “patio table” despite the phrase “patio table” not occurring in the input data 114); and generating, through the preset wallpaper generation model, new icons corresponding to the candidate wallpaper (Paragraph [0089]: he third panel 414 includes a user interface element 420 that is manipulable to specify an aspect ratio for the output images 124, and a user interface element 422 that is manipulable to specify one or more styles for the output images 124. The styles that are specifiable by the user interface element 422 include, for example, a degree of realism for the output images 124 (e.g., photorealistic images or cartoon images), color schemes for the output images 124, and the like. The third panel 414 additionally includes a user interface element 424 that is selectable to instruct the background generation system 112 to generate a new output image 124 in accordance with the updated parameters specified via the third panel 414. In response to receiving user input selecting the user interface element 424, for instance, the client device 204 communicates the updated parameters to the background generation system 112, which generates a new output image 124 in accordance with the techniques discussed herein based on the updated parameters. The new output image 124 is communicated back to the client device 204, which renders the new output image 124 for display in the first panel 410 and/or the second panel 412).
Regarding claim 11, Jain discloses the method for wallpaper setting of claim 1, wherein comparing the at least two wallpaper setting conditions with the information tags of the wallpapers in the wallpaper database and matching out, from the wallpaper database, the candidate wallpapers each of which has information tags comprising the at least two wallpaper setting conditions comprise: obtaining user use information of the electronic device (Fig. 2; Fig. 3; Paragraph [0053]: FIG. 3 depicts a system 300 in an example implementation showing operation of a background generation system 112 to generate an output image 124 based on visual saliency. As shown, the background generation system 112 receives the input data 114, e.g., responsive to receiving the notification 230. As previously mentioned, the input data 114 includes the user data 210 of the user 232 requesting to access a publication 212, the images 116 and the textual information 120 of the publication 212 being accessed, and/or image tags 234 extracted from the images 116 by the tag extraction model 236); and comparing the at least two wallpaper setting conditions with the information tags of the wallpapers in the wallpaper database, and matching out, from the wallpaper database, candidate wallpapers, wherein information tags of each of the candidate wallpapers comprise the at least two wallpaper setting conditions and each of the candidate wallpapers matches the user use information (Paragraphs [0054]-[0056]: the input data 114 (e.g., the textual information 120, the user data 210, and/or the image tags 234) is provided as input to a feature extraction module 302. The feature extraction module 302 includes a knowledge graph 304 (also known as a semantic network), which is a network of entities 306 that includes associations 308 that connect different entities 306 and describe how the connected entities 306 are related. By way of example, an entity 306 “umbrella” and the entity 306 “picnics” are connected by a usage association 308 in the knowledge graph 304, i.e., the umbrella is “used for” picnics. Broadly, the feature extraction module 302 queries the knowledge graph 304 with the input data 114, and employs an entity extraction technique to identify entities 306 from the knowledge graph 304 that correspond to the input data 114… Using the entity extraction technique, the feature extraction module 302 outputs features 310 including one or more of the entities 306 that match contiguous portions of text in the input data 114 and/or one or more additional entities 306 to which the matching entities 306 are connected in the knowledge graph 304. In an example, the feature extraction module 302 identifies the term “umbrella” that is present as an entity 306 in the knowledge graph 304, and in the input data 114. Further, the entity 306 “umbrella” is connected via an association 308 to the entity 306 “patio table.” In this example, the feature extraction module 302 outputs, as the features 310, the entity 306 “umbrella” and the entity 306 “patio table” despite the phrase “patio table” not occurring in the input data 114).
Regarding claim 14, Jain discloses the method for wallpaper setting of claim 1, wherein obtaining the at least two wallpaper setting conditions comprises: obtaining a phrase and/or a segment input by a user (Paragraph [0003]: image tags, the textual information, and/or the extracted features are provided as input to a generative text model, which is trained using machine learning to transform unformatted text into text prompts that inform a generative image model how to generate a background image for the object. Accordingly, the generative text model generates a prompt and provides the prompt as input to a generative image model, which is trained using machine learning to generate background images based on the prompt. Upon receiving the prompt and an image of the multiple images, the generative image model generates a background image based on the prompt. Further, an output image is generated by incorporating the object (as depicted in one of the multiple images) into the background image; Paragraph [0056]: Using the entity extraction technique, the feature extraction module 302 outputs features 310 including one or more of the entities 306 that match contiguous portions of text in the input data 114 and/or one or more additional entities 306 to which the matching entities 306 are connected in the knowledge graph 304. In an example, the feature extraction module 302 identifies the term “umbrella” that is present as an entity 306 in the knowledge graph 304, and in the input data 114. Further, the entity 306 “umbrella” is connected via an association 308 to the entity 306 “patio table.” In this example, the feature extraction module 302 outputs, as the features 310, the entity 306 “umbrella” and the entity 306 “patio table” despite the phrase “patio table” not occurring in the input data 114); and converting, through a preset language model, the phrase and/or the segment input by the user into the at least two wallpaper setting conditions (Paragraph [0064]: the generative text model 316 is a pre-trained large language model that is fine-tuned using supervised learning to generate prompts 318 that are suitable for the background generation task. Any one of a variety of public or proprietary large language models are employable by the background generation system 112, examples of which include T5, GPT-3, GPT-3.5, and GPT-4 models, to name just a few. In particular, the generative text model 316 is trained on pairs of training data including sets of curated input data and corresponding target prompts. During the training process, the generative text model 316 is conditioned on the curated input data. Further, the generative text model 316 outputs a prompt 318 in accordance with the described techniques).
Regarding claim 15, the limitations of this claim substantially correspond to the limitations of claim 1 (except for the display, processor, and memory, which are disclosed by Jain, Fig. 7); thus they are rejected on similar grounds.
Regarding claim 18, the limitations of this claim substantially correspond to the limitations of claim 15; thus they are rejected on similar grounds.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 2, 16, and 19 are rejected under 35 U.S.C. 103 as being unpatentable over Jain, in view of Clatworthy et al. (US Pub. 2007/0147654), hereinafter Clatworthy.
Regarding claim 2, Jain discloses the method for wallpaper setting of claim 1, wherein before obtaining the at least two wallpaper setting conditions, the method for wallpaper setting further comprises: obtaining a wallpaper generation instruction, and obtaining at least two prompts according to the wallpaper generation instruction (Paragraph [0056]: the feature extraction module 302 outputs features 310 including one or more of the entities 306 that match contiguous portions of text in the input data 114 and/or one or more additional entities 306 to which the matching entities 306 are connected in the knowledge graph 304. In an example, the feature extraction module 302 identifies the term “umbrella” that is present as an entity 306 in the knowledge graph 304, and in the input data 114. Further, the entity 306 “umbrella” is connected via an association 308 to the entity 306 “patio table.” In this example, the feature extraction module 302 outputs, as the features 310, the entity 306 “umbrella” and the entity 306 “patio table” despite the phrase “patio table” not occurring in the input data 114; Paragraph [0088]: the third panel 414 includes a user interface element 418 that is manipulable to update the attributes used by the background generation system 112 to condition the generative text model 316 for prompt generation. The user interface element 418 includes attributes (e.g., “Umbrella,” “Picnic,” “Outdoors,”) that have been identified automatically by the background generation system 112. These attributes, for example, include one or more of the features 310 extracted from the input data 114, one or more unprocessed portions of the input data 114, one or more of the image tags 234, and/or the predicted usage 314. Each of the automatically identified attributes are selectable (e.g., as illustrated by the “x” proximate the attributes) to remove a respective attribute from consideration by the generative text model 316. Moreover, the user interface element 418 includes a search bar that is manipulable (e.g., via text input) to specify additional attributes for consideration by the generative text model 316); generating, through the preset wallpaper generation model, wallpapers according to a wallpaper generation sentence obtained from the at least two prompts, wherein the wallpaper generation sentence comprises a plurality of restrictive conditions based on the at least two prompts (Fig. 3; Paragraphs [0060]-[0062]: a generative text model 316 receives, as input, the features 310, the predicted usage 314, and/or the unprocessed input data 114. In particular, the unprocessed input data 114 received by the generative text model 316 includes the image tags 234, the textual information 120, and/or the user data 210. Broadly, the generative text model 316 is a machine learning text-to-text model that is trained to take unformatted text as input, and output a prompt 318 that is suitable for background image generation by a generative image model 320. Notably, the generative text model 316 and the generative image model 320 collectively form the generative AI models…These sources of input data 114 are structured in the sense that the textual data is associated with a topic or context. By way of example, an object usage 214 that is included as part of the publication 212 includes text that is within the context of a usage of the object 118 that is the subject of the publication 212; Paragraphs [0065]-[0066]: generative text model 316 is trained by comparing the generated prompt 318 to the corresponding target prompt. By way of example, the generated prompt 318 and the target prompt are converted to vectors using a text vectorization technique (e.g., Term Frequency-Inverse Document Frequency (TF-IDF), Word2Vec, and GloVe), and compared using a similarity metric (e.g., cosine similarity or Euclidean distance). Parameters of the generative text model 316 are iteratively updated based on degrees of similarity between the generated prompt 318 and the target prompt on different pairs of training data…the generative image model 320 receives the prompt 318 and the input data 114 as input. In particular, the input data 114 provided to the generative image model 320 is an image 116 depicting the object 118 for which the background image 126 is to be generated. Broadly, the generative image model 320 is a machine learning text-to-image model that is trained to receive a prompt 318 and an image 116 as input, and generate an output image 124 having a new background that is based on the prompt 318. To do so, the generative image model 320 generates a background image 126 based on the prompt 318. Further, the generative image model 320 replaces a background (as depicted in the image 116) with the background image 126, thereby incorporating the object (as depicted in the image 116) into the background image).
Jain does not explicitly disclose forming the wallpaper database based on the wallpapers generated through the preset wallpaper generation model, wherein information tags of each of the wallpapers generated according to the at least two prompts comprise at least part of the at least two prompts.
However, Clatworthy teaches generation of background images based on prompts (Paragraph [0078]), further comprising forming the wallpaper database based on the wallpapers generated through the preset wallpaper generation model, wherein information tags of each of the wallpapers generated according to the at least two prompts comprise at least part of the at least two prompts (Fig. 10; Paragraphs [0028]-[0036]: FIG. 10 is an example series of frames generated by the cinematic frame creation system using a custom database of character images and backgrounds, in accordance with an embodiment of the present invention… Examples of frames generated by the cinematic frame creation system 145 are shown in FIGS. 9 and 10. FIG. 9 illustrates two example assembled frames generated by the cinematic frame creation system 145, in accordance with two embodiments of the present invention. The first frame 901 is a two-shot and an over-the-shoulder shot and was created for a Television aspect ratio (1.33). The second frame 902 includes the same content (a two-shot and an over-the-shoulder shot) but object placement is adjusted for a wide-screen format. The second frame 902 has less headroom and a wider background is visible than the first frame 901. In both frames 901 and 902, the characters are distributed in cinematically pleasing composition based on variety of cinematic conventions mentioned above, e.g., headroom, ground space, horizon, edging, etc. FIG. 10 is an example series of three frames 1001, 1002 and 1003 generated by the cinematic frame creation system 145 using a custom database of character renderings and backgrounds, in accordance with an embodiment of the present invention). Clatworthy teaches that this will allow for saving time in image selection (Paragraph [0015]). Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to have modified Jain with the features above as taught by Clatworthy so as to save time in image selection as presented by Clatworthy.
Regarding claim 16, the limitations of this claim substantially correspond to the limitations of claim 2; thus they are rejected on similar grounds.
Regarding claim 19, the limitations of this claim substantially correspond to the limitations of claim 16; thus they are rejected on similar grounds.
Allowable Subject Matter
Claims 3, 8, 10, 12, 17, and 20 are objected to as being dependent upon a rejected base claim, but would be allowable if rewritten in independent form including all of the limitations of the base claim and any intervening claims.
Claims 3, 17, and 20 are allowable over the prior art of record since the cited references taken individually or in combination fails to particularly disclose or suggest a method, device, or medium, comprising: eliminating a wallpaper whose image content and/or information tags meet a preset elimination rule from the wallpapers generated through the preset wallpaper generation model; and adding remaining wallpapers in the wallpapers generated through the preset wallpaper generation model to the wallpaper database, as presented in the environment of the remaining limitations of claim 3 (and substantially similar limitations in each of claims 17 and 20). It is noted that the closest prior art, Jain, shows the limitations of claim 2 (and 16 and 19, respectively), wherein forming the wallpaper database based on the wallpapers generated through the preset wallpaper generation model comprises: parsing each of the wallpapers generated through the preset wallpaper generation model, to obtain image content; obtaining information tags of each of the wallpapers based on the image content. However, the Jain fails to disclose or suggest eliminating a wallpaper whose image content and/or information tags meet a preset elimination rule from the wallpapers generated through the preset wallpaper generation model; and adding remaining wallpapers in the wallpapers generated through the preset wallpaper generation model to the wallpaper database.
Claim 8 is allowable over the prior art of record since the cited references taken individually or in combination fails to particularly disclose or suggest a method comprising: generating, through the preset wallpaper generation model according to the semantic association result, a candidate wallpaper that matches the layout of the icons on the display interface, as presented in the environment of the remaining limitations of claim 8. It is noted that the closest prior art, Jain, shows the limitations of claim 5 and wherein performing at least semantic association on the at least two wallpaper setting conditions and directly invoking the preset wallpaper generation model to generate, according to the semantic association result, the new wallpaper comprise: identifying a layout of icons on a display interface currently displayed on the display of the electronic device; and performing semantic association on the at least two wallpaper setting conditions. However, the Jain fails to disclose or suggest generating, through the preset wallpaper generation model according to the semantic association result, a candidate wallpaper that matches the layout of the icons on the display interface.
Claim 10 is allowable over the prior art of record since the cited references taken individually or in combination fails to particularly disclose or suggest a method comprising: performing dynamic transition processing on two adjacent static images in each set of static images based on the target element, to obtain one dynamic wallpaper, wherein a number of times that dynamic transition processing is performed on any two adjacent static images in each set of static images is at least once, and one corresponding transition image is obtained through each dynamic transition processing, as presented in the environment of the remaining limitations of claim 10. It is noted that the closest prior art, Jain, shows the limitations of claim 2 and wherein the wallpapers generated through the preset wallpaper generation model comprise a dynamic wallpaper, and generating, through the preset wallpaper generation model, the wallpapers according to the wallpaper generation sentence obtained from the at least two prompts comprises: obtaining a corresponding wallpaper generation sentence according to the at least two prompts; generating, through the preset wallpaper generation model according to the wallpaper generation sentence, at least one set of static images each comprising a target element, wherein the target element is associated with the wallpaper generation sentence, and each set of static images comprises at least two static images. However, the Jain fails to disclose or suggest performing dynamic transition processing on two adjacent static images in each set of static images based on the target element, to obtain one dynamic wallpaper, wherein a number of times that dynamic transition processing is performed on any two adjacent static images in each set of static images is at least once, and one corresponding transition image is obtained through each dynamic transition processing.
Claim 12 is allowable over the prior art of record since the cited references taken individually or in combination fails to particularly disclose or suggest a method comprising: counting a number of times that each of the candidate wallpapers is used; and displaying m candidate wallpapers with the lowest usage counts among all the candidate wallpapers on the display of the electronic device in an ascending order of usage counts, wherein m≥1, as presented in the environment of the remaining limitations of claim 12. It is noted that the closest prior art, Jain, shows the limitations of claim 1. However, the Jain fails to disclose or suggest counting a number of times that each of the candidate wallpapers is used; and displaying m candidate wallpapers with the lowest usage counts among all the candidate wallpapers on the display of the electronic device in an ascending order of usage counts, wherein m≥1.
Claim 13 is allowable over the prior art of record since the cited references taken individually or in combination fails to particularly disclose or suggest a method comprising: in response to a random wallpaper generation instruction, obtaining randomly one first-type wallpaper setting condition from the plurality of first-type wallpaper setting conditions, and obtaining randomly one second-type wallpaper setting condition from the plurality of second-type wallpaper setting conditions, as presented in the environment of the remaining limitations of claim 13. It is noted that the closest prior art, Jain, shows the limitations of claim 1. However, the Jain fails to disclose or suggest in response to a random wallpaper generation instruction, obtaining randomly one first-type wallpaper setting condition from the plurality of first-type wallpaper setting conditions, and obtaining randomly one second-type wallpaper setting condition from the plurality of second-type wallpaper setting conditions.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Ghosh et al. (US Pub. 2024/0104789) teaches image generation via text prompt.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MATTHEW D SALVUCCI whose telephone number is (571)270-5748. The examiner can normally be reached M-F: 7:30-4:00PT.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, XIAO WU can be reached at (571) 272-7761. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/MATTHEW SALVUCCI/Primary Examiner, Art Unit 2613
Prosecution Insights