A digital photo frame display intelligent adjustment system based on image content recognition

Abstract

The present invention discloses a digital photo frame display intelligent adjustment system based on image content recognition, comprising an acquisition module for acquiring the current display image of the digital photo frame; an identification module for performing image content recognition on the current display image to obtain first content information; and a first control module for determining whether the first content information is consistent with preset second content information. If it is determined that the first content information and the preset second content information are inconsistent, the system queries a database based on the second content information to obtain a target image including the second content information, and controls the digital photo frame to display the target image. Advantageous Effects: By comparing the first content information with the second content information preset by the user, the display image can be changed in a timely manner, thereby solving the problem of single display mode of the digital photo frame, increasing the diversity of the display mode of the digital photo frame, and improving the intelligence of the digital photo frame.

Description

Technical Field

[0001] The present invention relates to the technical field of digital photo frames, and in particular to a digital photo frame display intelligent adjustment system based on image content recognition. Background Art

[0002] With the advent and popularity of digital cameras, digital photos have become the primary form of photography, stored on computers, portable hard drives, memory cards, and other media. Digital photo frames are digital products that receive and display digital photos stored on computers, cameras, camcorders, SD cards, mobile devices, resource websites, and other devices. Currently, digital photo frames on the market are relatively simple and lack intelligence. They cannot automatically identify the content of displayed images, let alone automatically change the display based on the content of the displayed image. The monotony and fixed nature of digital photo frames' displays is a problem that urgently needs to be addressed. Summary of the Invention

[0003] The present invention aims to at least partially address one of the technical problems in the above-mentioned technologies. To this end, the present invention provides a digital photo frame display intelligent adjustment system based on image content recognition, which compares first content information with second content information preset by the user and then promptly changes the displayed image. This solves the problem of single display mode in digital photo frames, increases the diversity of display modes in digital photo frames, and improves the intelligence of digital photo frames.

[0004] To achieve the above objectives, the present invention proposes a digital photo frame display intelligent adjustment system based on image content recognition, comprising:

[0005] An acquisition module, configured to acquire the current display image of the digital photo frame;

[0006] an identification module, configured to perform image content identification on the currently displayed image to obtain first content information;

[0007] The first control module is configured to determine whether the first content information is consistent with the preset second content information; when it is determined that the first content information is inconsistent with the preset second content information, query a database based on the second content information to obtain a target image including the second content information, and control the digital photo frame to display the target image.

[0008] Furthermore, it also includes:

[0009] A working state detection module, used for detecting whether the digital photo frame is in working state;

[0010] A human body detection module, used to detect whether there is a human body around the digital photo frame;

[0011] The second control module is connected to the working state detection module and the human body detection module respectively, and is used to turn off the digital photo frame when it is determined that the digital photo frame is in working state and there is no human body around the digital photo frame.

[0012] Furthermore, the working status detection module includes:

[0013] A brightness signal acquisition module, configured to acquire a brightness signal of the digital photo frame;

[0014] The first judgment module is configured to analyze the brightness signal to obtain a brightness value, determine whether the brightness value is greater than a preset brightness value, and determine that the digital photo frame is in a working state when it is determined that the brightness value is greater than the preset brightness value.

[0015] Furthermore, the human body detection module includes:

[0016] An infrared energy detection module, used to detect the infrared energy value of the environment surrounding the digital photo frame;

[0017] A second judgment module is used to judge whether the infrared energy value has a sudden change;

[0018] a third control module, configured to obtain an image of a mutation region when the second judgment module determines that a mutation has occurred in the infrared energy value, perform human body recognition on the image of the mutation region according to human body recognition technology, obtain a recognition result, and determine whether there is a human body around the digital photo frame according to the recognition result.

[0019] Furthermore, before performing human body recognition on the image of the sudden change region according to the human body recognition technology, the method further includes performing noise reduction processing on the image of the sudden change region.

[0020] Furthermore, it also includes:

[0021] An environmental image acquisition module, used to acquire an environmental image around the digital photo frame;

[0022] Ambient color temperature acquisition module, used for:

[0023] Performing image segmentation processing on the environment image to obtain a plurality of sub-environment images;

[0024] Obtaining the RGB color value of each pixel in the sub-environment image, wherein the RGB color value includes an R channel value, a G channel value, and a B channel value;

[0025] The R channel average value is calculated based on the R channel value of each pixel in the sub-environment image;

[0026] The G channel average value is calculated based on the G channel value of each pixel in the sub-environment image;

[0027] The B channel average value is calculated based on the B channel value of each pixel in the sub-environment image;

[0028] Calculating a ratio of the R channel average value to the G channel average value as a first ratio;

[0029] Calculating a ratio of the B channel average value to the G channel average value as a second ratio;

[0030] Generate a coordinate point according to the first ratio and the second ratio, mark the coordinate point on a preset color temperature map to obtain a marked point, obtain a color temperature value corresponding to the marked point, and use the color temperature value as the color temperature value of the sub-environment image;

[0031] Comparing the color temperature values of the plurality of sub-environment images with a preset threshold value, selecting the sub-environment images whose color temperature values are greater than or equal to the preset threshold value, and generating a first set;

[0032] Filtering sub-environment images with color temperature values smaller than a preset threshold and generating a second set;

[0033] Calculating a first color temperature average value according to the color temperature values of the sub-environment images included in the first set;

[0034] Calculating a second color temperature average value according to the color temperature values of the sub-environment images included in the second set;

[0035] Counting a first number of sub-environment images included in the first set;

[0036] Counting a second number of sub-environment images included in the second set;

[0037] calculating a sum of the first quantity and the second quantity;

[0038] calculating a third ratio of the first quantity to the sum, and using the third ratio as a first weight coefficient;

[0039] calculating a fourth ratio of the second quantity to the sum, and using the fourth ratio as a second weight coefficient;

[0040] Calculate an ambient color temperature value according to the first color temperature average value, the second color temperature average value, the first weight coefficient, and the second weight coefficient;

[0041] A current color temperature value acquisition module, used to obtain the current color temperature value of the digital photo frame;

[0042] a fourth control module, configured to query a preset ambient color temperature value-target color temperature value table according to the ambient color temperature value, obtain a corresponding target color temperature value, calculate a difference between the current color temperature value and the target color temperature value, use the difference as a color temperature adjustment coefficient, and adjust the current color temperature value of the digital photo frame according to the color temperature adjustment coefficient.

[0043] Furthermore, the identification module includes:

[0044] A grayscale module, configured to perform image preprocessing on the current display image of the digital photo frame to obtain a corresponding grayscale image;

[0045] Image processing module for:

[0046] Obtaining a first grayscale value of each pixel in the gradient image, and screening out the largest first grayscale value;

[0047] Performing a convolution operation on the first grayscale value of each pixel with a preset horizontal gradient operator to obtain a horizontal gradient value of each pixel, and obtaining a first absolute value of the horizontal gradient value of each pixel;

[0048] Filter out the largest first absolute value and the smallest first absolute value, and calculate the first difference;

[0049] Subtracting the first absolute value of each pixel from the minimum first absolute value to obtain a plurality of second difference values;

[0050] Calculating the ratio of a plurality of second difference values to the first difference value respectively to obtain a second grayscale value of each pixel;

[0051] Performing a convolution operation on the first grayscale value of each pixel with a preset vertical gradient operator to obtain a vertical gradient value of each pixel, and obtaining a second absolute value of the vertical gradient value of each pixel;

[0052] Filter out the largest second absolute value and the smallest second absolute value, and calculate the third difference;

[0053] subtracting the second absolute value of each pixel from the minimum second absolute value to obtain a plurality of fourth difference values;

[0054] Calculating the ratio of a plurality of fourth difference values to the second difference value respectively to obtain a third grayscale value of each pixel;

[0055] Performing weighted summation processing on the second grayscale value and the third grayscale value of each pixel to obtain a target grayscale value of each pixel, determining a pixel value of each pixel based on the target grayscale value of each pixel, and generating a gradient image based on the pixel value of each pixel;

[0056] Model training module, used for:

[0057] Building a content recognition model;

[0058] Obtaining a sample image set, inputting the sample image set into the content recognition model and performing iterative training according to a preset loss function;

[0059] During the training of the content recognition model, obtaining the number of trained iterations of the content recognition model and determining whether the number of trained iterations is greater than a preset number of iterations; when it is determined that the number of trained iterations is greater than the preset number of iterations, adjusting the parameters of the preset loss function to obtain a target loss function;

[0060] The content recognition model is iteratively trained according to the target loss function until the number of iterations is completed to obtain a trained content recognition model;

[0061] The fifth control module is configured to input the gradient image into a trained content recognition model and output corresponding first content information.

[0062] Furthermore, it also includes:

[0063] A contrast acquisition module, configured to acquire the contrast of the digital photo frame;

[0064] The sixth control module is configured to determine whether the contrast is within a preset contrast range, and adjust the contrast when it is determined that the contrast is not within the preset contrast range.

[0065] Furthermore, the determining whether the infrared energy value has a sudden change includes:

[0066] Obtain infrared energy values at two adjacent time points, calculate the infrared energy difference, determine whether the infrared energy difference is within a preset infrared energy difference range, and if it is determined that the infrared energy difference is not within the set infrared energy difference range, determine that the infrared energy value has undergone a mutation.

[0067] Furthermore, the performing denoising on the mutation region image includes inputting the mutation region image into a pre-trained denoising model and outputting a denoised mutation region image.

[0068] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0070] Figure 1 A block diagram of a digital photo frame display intelligent adjustment system based on image content recognition according to a first embodiment of the present invention;

[0071] Figure 2 A block diagram of a digital photo frame display intelligent adjustment system based on image content recognition according to a second embodiment of the present invention;

[0072] Figure 3 FIG. 4 is a block diagram of a digital photo frame display intelligent adjustment system based on image content recognition according to a third embodiment of the present invention. DETAILED DESCRIPTION

[0073] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0074] Reference below Figures 1 to 3 The following describes an intelligent digital photo frame display adjustment system based on image content recognition proposed in an embodiment of the present invention.

[0075] like Figure 1 As shown, a digital photo frame display intelligent adjustment system based on image content recognition includes:

[0076] An acquisition module, configured to acquire the current display image of the digital photo frame;

[0077] an identification module, configured to perform image content identification on the currently displayed image to obtain first content information;

[0078] The first control module is configured to determine whether the first content information is consistent with the preset second content information; when it is determined that the first content information is inconsistent with the preset second content information, query a database based on the second content information to obtain a target image including the second content information, and control the digital photo frame to display the target image.

[0079] The working principle of the above scheme is as follows: an acquisition module is used to acquire the current display image of the digital photo frame; an identification module is used to perform image content identification on the current display image to obtain first content information; a first control module is used to determine whether the first content information is consistent with preset second content information. When it is determined that the first content information is inconsistent with the preset second content information, a database is queried according to the second content information to obtain a target image including the second content information, and the digital photo frame is controlled to display the target image; wherein the content information of the image includes mountains, water, trees, people, flowers, birds, etc., and an image may contain multiple content information; for example, if the first content information is people and flowers, and the second content information is people and water, then the first content information is inconsistent with the preset second content information; according to the second content information, the database is queried, that is, according to the database of people and water, a target image including people and water is obtained. The target image may be one or multiple images. When the target image is multiple images, the target images are displayed in a loop.

[0080] The beneficial effects of the above solution are as follows: the first content information is compared with the second content information preset by the user, and the displayed image is replaced in time, which solves the single display mode of the digital photo frame, increases the diversity of the display mode of the digital photo frame, and improves the intelligence of the digital photo frame.

[0081] According to some embodiments of the present invention, the further comprising:

[0082] A working state detection module, used for detecting whether the digital photo frame is in working state;

[0083] A human body detection module, used to detect whether there is a human body around the digital photo frame;

[0084] The second control module is connected to the working state detection module and the human body detection module respectively, and is used to turn off the digital photo frame when it is determined that the digital photo frame is in working state and there is no human body around the digital photo frame.

[0085] The working principle of the above scheme is as follows: the working status detection module is used to detect whether the digital photo frame is in the working state; the human body detection module is used to detect whether there is a human body around the digital photo frame; the second control module is used to turn off the digital photo frame when it is determined that the digital photo frame is in the working state and there is no human body around the digital photo frame.

[0086] The beneficial effect of the above solution is that when it is determined that the digital photo frame is in working state and there is no human body around the digital photo frame, the digital photo frame is turned off, thereby saving resources and improving user experience.

[0087] According to some embodiments of the present invention, the working status detection module includes:

[0088] A brightness signal acquisition module, configured to acquire a brightness signal of the digital photo frame;

[0089] The first judgment module is configured to analyze the brightness signal to obtain a brightness value, determine whether the brightness value is greater than a preset brightness value, and determine that the digital photo frame is in a working state when it is determined that the brightness value is greater than the preset brightness value.

[0090] The working principle of the above scheme is as follows: a brightness signal acquisition module is used to obtain the brightness signal of the digital photo frame; a first judgment module is used to analyze the brightness signal to obtain a brightness value, and judge whether the brightness value is greater than a preset brightness value. When it is determined that the brightness value is greater than the preset brightness value, it is determined that the digital photo frame is in a working state.

[0091] The beneficial effect of the above solution is that whether the digital photo frame is in a working state is judged according to the brightness value of the digital photo frame, so that the final result is more accurate.

[0092] According to some embodiments of the present invention, the human body detection module includes:

[0093] An infrared energy detection module, used to detect the infrared energy value of the environment surrounding the digital photo frame;

[0094] A second judgment module is used to judge whether the infrared energy value has a sudden change;

[0095] a third control module, configured to obtain an image of a mutation region when the second judgment module determines that a mutation has occurred in the infrared energy value, perform human body recognition on the image of the mutation region according to human body recognition technology, obtain a recognition result, and determine whether there is a human body around the digital photo frame according to the recognition result.

[0096] The working principle of the above scheme is as follows: an infrared energy detection module is used to detect the infrared energy value of the environment around the digital photo frame; a second judgment module is used to judge whether the infrared energy value has undergone a mutation; a third control module is used to obtain an image of the mutation area when the second judgment module determines that the infrared energy value has undergone a mutation, and perform human body recognition on the image of the mutation area based on human body recognition technology to obtain a recognition result, and determine whether there is a human body around the digital photo frame based on the recognition result.

[0097] The beneficial effects of the above scheme are as follows: the steps for detecting the human body are specifically described, a preliminary detection is performed based on infrared energy, and then a secondary detection is performed based on the image, thereby improving the accuracy of the detection results.

[0098] According to some embodiments of the present invention, before performing human body recognition on the image of the sudden change region according to the human body recognition technology, the method further includes performing noise reduction processing on the image of the sudden change region.

[0099] The working principle of the above solution is: before performing human body recognition on the image of the mutation area according to the human body recognition technology, it also includes performing noise reduction processing on the image of the mutation area.

[0100] The beneficial effect of the above solution is that the noise reduction process is performed on the image of the mutation region to make the image of the mutation region clearer after noise reduction.

[0101] like Figure 2 As shown, according to some embodiments of the present invention, it also includes:

[0102] An environmental image acquisition module, used to acquire an environmental image around the digital photo frame;

[0103] Ambient color temperature acquisition module, used for:

[0104] Performing image segmentation processing on the environment image to obtain a plurality of sub-environment images;

[0105] Obtaining the RGB color value of each pixel in the sub-environment image, wherein the RGB color value includes an R channel value, a G channel value, and a B channel value;

[0106] The R channel average value is calculated based on the R channel value of each pixel in the sub-environment image;

[0107] The G channel average value is calculated based on the G channel value of each pixel in the sub-environment image;

[0108] The B channel average value is calculated based on the B channel value of each pixel in the sub-environment image;

[0109] Calculating a ratio of the R channel average value to the G channel average value as a first ratio;

[0110] Calculating a ratio of the B channel average value to the G channel average value as a second ratio;

[0111] Generate a coordinate point according to the first ratio and the second ratio, mark the coordinate point on a preset color temperature map to obtain a marked point, obtain a color temperature value corresponding to the marked point, and use the color temperature value as the color temperature value of the sub-environment image;

[0112] Comparing the color temperature values of the plurality of sub-environment images with a preset threshold value, selecting the sub-environment images whose color temperature values are greater than or equal to the preset threshold value, and generating a first set;

[0113] Filtering sub-environment images with color temperature values smaller than a preset threshold and generating a second set;

[0114] Calculating a first color temperature average value according to the color temperature values of the sub-environment images included in the first set;

[0115] Calculating a second color temperature average value according to the color temperature values of the sub-environment images included in the second set;

[0116] Counting a first number of sub-environment images included in the first set;

[0117] Counting a second number of sub-environment images included in the second set;

[0118] calculating a sum of the first quantity and the second quantity;

[0119] calculating a third ratio of the first quantity to the sum, and using the third ratio as a first weight coefficient;

[0120] calculating a fourth ratio of the second quantity to the sum, and using the fourth ratio as a second weight coefficient;

[0121] Calculate an ambient color temperature value according to the first color temperature average value, the second color temperature average value, the first weight coefficient, and the second weight coefficient;

[0122] A current color temperature value acquisition module, used to obtain the current color temperature value of the digital photo frame;

[0123] a fourth control module, configured to query a preset ambient color temperature value-target color temperature value table according to the ambient color temperature value, obtain a corresponding target color temperature value, calculate a difference between the current color temperature value and the target color temperature value, use the difference as a color temperature adjustment coefficient, and adjust the current color temperature value of the digital photo frame according to the color temperature adjustment coefficient.

[0124] The working principle of the above scheme is as follows: the environmental image acquisition module is used to acquire the environmental image around the digital photo frame; the environmental color temperature value acquisition module is used to perform image segmentation processing on the environmental image to obtain a plurality of sub-environmental images; the RGB color value of each pixel in the sub-environmental image is obtained, and the RGB color value includes an R channel value, a G channel value and a B channel value; the R channel average value is calculated according to the R channel value of each pixel in the sub-environmental image; the G channel average value is calculated according to the G channel value of each pixel in the sub-environmental image; the B channel average value is calculated according to the B channel value of each pixel in the sub-environmental image; the ratio of the R channel average value to the G channel average value is calculated as a first ratio; the ratio of the B channel average value to the G channel average value is calculated as a second ratio; a coordinate point is generated according to the first ratio and the second ratio, and a mark is made on a preset color temperature map according to the coordinate point to obtain a marked point, and the color temperature value corresponding to the marked point is obtained and used as the color temperature value of the sub-environmental image;

[0125] Generally speaking, if a pixel point is located to the right or below the position of the preset color temperature map, it means that the color temperature of this pixel point is lower; relatively, if a pixel point is located to the left or above the position of the preset color temperature map, it means that the color temperature of this pixel point is higher; compare the color temperature values of several sub-environment images with preset thresholds respectively, filter out sub-environment images with color temperature values greater than or equal to the preset threshold, and generate a first set; filter out sub-environment images with color temperature values less than the preset threshold, and generate a second set; calculate the first color temperature average value based on the color temperature values of the sub-environment images included in the first set; calculate the second color temperature average value based on the color temperature values of the sub-environment images included in the second set; count the first number of sub-environment images included in the first set; count the second number of sub-environment images included in the second set; calculate the sum of the first number and the second number; calculate the third ratio of the first number to the sum, and use the third ratio as the first weight coefficient; calculate the fourth ratio of the second number to the sum, and use the fourth ratio as the weight coefficient is a second weight coefficient; an ambient color temperature value is calculated according to the first color temperature average value, the second color temperature average value, the first weight coefficient and the second weight coefficient; a current color temperature value acquisition module is used to obtain the current color temperature value of the digital photo frame; a fourth control module is used to query a preset ambient color temperature value-target color temperature value table according to the ambient color temperature value, obtain a corresponding target color temperature value, calculate the difference between the current color temperature value and the target color temperature value, use the difference as the color temperature adjustment coefficient, and adjust the color temperature value according to the color temperature adjustment coefficient. The adjustment coefficient is used to adjust the current color temperature value of the digital photo frame; for example, if the calculated ambient color temperature value is 2000K, the preset ambient color temperature value-target color temperature value table is queried according to the ambient color temperature value, and the target color temperature is 2300K. The current color temperature value of the digital photo frame is 2500K, and the difference between the current color temperature value and the target color temperature value is calculated, that is, 2500K-2300K=200K, that is, 200K is the adjustment coefficient, and the current color temperature value of the digital photo frame is adjusted up by 200K.

[0126] The beneficial effects of the above scheme: the color temperature value of the digital photo frame is an important display parameter of the digital photo frame. Color temperature is an important parameter used to characterize the color characteristics of light. The lower the color temperature, the redder the light color, and vice versa, the bluer the light color. The overall color temperature of the image can characterize the overall feeling of the image to the viewer. Too high or too low a color temperature will have a bad effect on the user's eyes. The color temperature of the environment around the digital photo frame is the main factor affecting the viewer. Therefore, it is necessary to adjust the color temperature of the digital photo frame in real time and dynamically according to the color temperature value of the surrounding environment. This scheme provides a fully automatic, fast and highly accurate adjustment method. First, the environmental image around the digital photo frame is obtained. Obtaining the environmental image is a necessary prerequisite for detecting the environmental color temperature value; then, the coordinate point is obtained according to the first ratio and the second ratio of each sub-environmental image, and the coordinate point is marked on the preset color temperature map according to the coordinate point, so that The obtained color temperature value is more accurate; the color temperature values of several sub-environmental images are comprehensively processed, and finally the ambient color temperature value of the obtained ambient image is made more accurate, and then a preset ambient color temperature value-target color temperature value table is queried according to the ambient color temperature value, the difference between the current color temperature value and the target color temperature value is calculated, the difference is used as a color temperature adjustment coefficient, and the current color temperature value of the digital photo frame is adjusted according to the color temperature adjustment coefficient; the current color temperature value of the digital photo frame is dynamically adjusted so that the current color temperature value of the digital photo frame is always in an optimal state, thereby ensuring that the user's eyes are not harmed and improving the user experience.

[0127] like Figure 3 As shown, according to some embodiments of the present invention, the identification module includes:

[0128] A grayscale module, configured to perform image preprocessing on the current display image of the digital photo frame to obtain a corresponding grayscale image;

[0129] Image processing module for:

[0130] Obtaining a first grayscale value of each pixel in the gradient image, and screening out the largest first grayscale value;

[0131] Performing a convolution operation on the first grayscale value of each pixel with a preset horizontal gradient operator to obtain a horizontal gradient value of each pixel, and obtaining a first absolute value of the horizontal gradient value of each pixel;

[0132] Filter out the largest first absolute value and the smallest first absolute value, and calculate the first difference;

[0133] Subtracting the first absolute value of each pixel from the minimum first absolute value to obtain a plurality of second difference values;

[0134] Calculating the ratio of a plurality of second difference values to the first difference value respectively to obtain a second grayscale value of each pixel;

[0135] Performing a convolution operation on the first grayscale value of each pixel with a preset vertical gradient operator to obtain a vertical gradient value of each pixel, and obtaining a second absolute value of the vertical gradient value of each pixel;

[0136] Filter out the largest second absolute value and the smallest second absolute value, and calculate the third difference;

[0137] subtracting the second absolute value of each pixel from the minimum second absolute value to obtain a plurality of fourth difference values;

[0138] Calculating the ratio of a plurality of fourth difference values to the second difference value respectively to obtain a third grayscale value of each pixel;

[0139] Performing weighted summation processing on the second grayscale value and the third grayscale value of each pixel to obtain a target grayscale value of each pixel, determining a pixel value of each pixel based on the target grayscale value of each pixel, and generating a gradient image based on the pixel value of each pixel;

[0140] Model training module, used for:

[0141] Building a content recognition model;

[0142] Obtaining a sample image set, inputting the sample image set into the content recognition model and performing iterative training according to a preset loss function;

[0143] During the training of the content recognition model, obtaining the number of trained iterations of the content recognition model and determining whether the number of trained iterations is greater than a preset number of iterations; when it is determined that the number of trained iterations is greater than the preset number of iterations, adjusting the parameters of the preset loss function to obtain a target loss function;

[0144] The content recognition model is iteratively trained according to the target loss function until the number of iterations is completed to obtain a trained content recognition model;

[0145] The fifth control module is configured to input the gradient image into a trained content recognition model and output corresponding first content information.

[0146] The working principle of the above scheme is as follows: the grayscale module is used to perform image preprocessing on the current display image of the digital photo frame to obtain a corresponding grayscale image; the image processing module is used to obtain the first grayscale value of each pixel in the gradient image and screen out the largest first grayscale value; the first grayscale value of each pixel is convolved with a preset horizontal gradient operator to obtain the horizontal gradient value of each pixel, and the first absolute value of the horizontal gradient value of each pixel is obtained; the largest first absolute value and the smallest first absolute value are screened out, and the first difference is calculated; the first absolute value of each pixel is respectively Subtract the first absolute value from the minimum to obtain several second differences; calculate the ratio of the second differences to the first difference to obtain the second grayscale value of each pixel; perform a convolution operation on the first grayscale value of each pixel with a preset vertical gradient operator to obtain the vertical gradient value of each pixel, and obtain the second absolute value of the vertical gradient value of each pixel; filter out the largest second absolute value and the smallest second absolute value, and calculate the third difference; subtract the second absolute value of each pixel from the minimum second absolute value to obtain several fourth differences; calculate a ratio of a plurality of fourth difference values to the second difference value to obtain a third grayscale value for each pixel; performing a weighted summation process on the second grayscale value and the third grayscale value of each pixel to obtain a target grayscale value for each pixel; determining a pixel value for each pixel based on the target grayscale value of each pixel; and generating a gradient image based on the pixel value of each pixel; wherein the target grayscale value for each pixel is the pixel value of the gradient image; a model training module, configured to: establish a content recognition model; obtain a sample image set, input the sample image set into the content recognition model, and iteratively train the model according to a preset loss function; during the training of the content recognition model, obtain a number of trained iterations of the content recognition model, determine whether the number of trained iterations is greater than a preset number of iterations, and if it is determined that the number of trained iterations is greater than the preset number of iterations, adjust the parameters of the preset loss function to obtain a target loss function; iteratively train the content recognition model according to the target loss function until the number of iterations is completed, thereby obtaining a trained content recognition model; and a fifth control module, configured to input the gradient image into the trained content recognition model and output corresponding first content information.

[0147] The beneficial effects of the above scheme are as follows: image content recognition of the currently displayed image is of paramount importance, and this scheme provides a method for accurately performing image content recognition; first, the currently displayed image is converted into a gradient image, the features of the gradient image are more obvious, content recognition is performed based on the gradient image, and the accuracy of the final recognition result is increased, and the method of this scheme is used to perform gradient image recognition; since different gradient values correspond to different grayscale values, different gradient values will not be regarded as the same during image processing, thereby improving the accuracy of image content recognition; training the content recognition model, and in the process of training the content recognition model, dynamically adjusting the target loss function used for each iterative training based on the number of iterations, which can better fit the distribution of complex image samples and reduce the number of sample images with intermediate probability value distributions, thereby increasing the recall rate of samples while ensuring the accuracy of the convolutional neural network recognition results, and inputting the gradient image into the trained content recognition model training, so that the output first content information is more accurate, thereby improving the accuracy of judging whether the first content information is consistent with the second content information.

[0148] According to some embodiments of the present invention, the further comprising:

[0149] A contrast acquisition module, configured to acquire the contrast of the digital photo frame;

[0150] The sixth control module is configured to determine whether the contrast is within a preset contrast range, and adjust the contrast when it is determined that the contrast is not within the preset contrast range.

[0151] The working principle of the above scheme is: the contrast acquisition module is used to obtain the contrast of the digital photo frame; the sixth control module is used to determine whether the contrast is within a preset contrast range, and when it is determined that the contrast is not within the preset contrast range, adjust the contrast.

[0152] Beneficial effects of the above solution: Contrast is also a precise parameter of the digital photo frame. By adjusting the current contrast, the user can see more clearly and ensure the user's viewing experience.

[0153] According to some embodiments of the present invention, determining whether a sudden change occurs in the infrared energy value includes:

[0154] Obtain infrared energy values at two adjacent time points, calculate the infrared energy difference, determine whether the infrared energy difference is within a preset infrared energy difference range, and if it is determined that the infrared energy difference is not within the set infrared energy difference range, determine that the infrared energy value has undergone a mutation.

[0155] The working principle of the above scheme is: obtain the infrared energy values of two adjacent time points, calculate the infrared energy difference, judge whether the infrared energy difference is within the preset infrared energy difference range, and if it is determined that the infrared energy difference is not within the set infrared energy difference range, determine that the infrared energy value has undergone a mutation.

[0156] The beneficial effect of the above solution is that it detects whether the infrared energy value has a sudden change based on the infrared energy values at two adjacent time points, so that the final detection result is more accurate.

[0157] According to some embodiments of the present invention, performing denoising on the mutation region image includes inputting the mutation region image into a pre-trained denoising model and outputting a denoised mutation region image.

[0158] The working principle of the above solution is as follows: the mutation region image is input into a pre-trained noise reduction model, and the noise-reduced mutation region image is output.

[0159] The beneficial effect of the above solution is that the image of the mutation area after noise reduction obtained according to the noise reduction model is clearer.

[0160] According to some embodiments of the present invention, the current display image of the digital photo frame is preprocessed to obtain a corresponding grayscale image, including

[0161] Convert the current display image from RGB color space to LAB color space;

[0162] Calculate the color metric K(x,y) of the pixel point (x,y) in the current display image, as shown in formula (1):

[0163]

[0164] Wherein, L(x,y) is the value of the pixel point (x,y) in the currently displayed image in the L channel; γ(x,y) is the hue value of the pixel point (x,y) in the currently displayed image; A(x,y) is the value of the pixel point (x,y) in the currently displayed image in the A channel; B(x,y) is the value of the pixel point (x,y) in the currently displayed image in the B channel;

[0165] According to the color metric K(x,y) of the pixel point (x,y) in the current display image, the grayscale value F(x,y) of the pixel point (x,y) in the current display image is calculated, as shown in formula (2):

[0166]

[0167] Repeat the above steps to obtain the grayscale value of each pixel in the currently displayed image, and then obtain the grayscale image of the currently displayed image.

[0168] The working principle and beneficial effects of the above scheme are as follows: when calculating the grayscale value F(x,y) of the pixel point (x,y), factors such as the value of the pixel point (x,y) in the L channel, the hue value of the pixel point (x,y), and the value of the pixel point (x,y) in the A channel are considered, so that the calculated grayscale value is more accurate, thereby ensuring the accuracy of the final grayscale image, wherein the color measurement specifically includes the brightness characteristics of the pixel points; the grayscale image obtained by the above method retains more of the overall result information of the original color image, and retains most of the contrast information of the original color image, solves the problem of partial grayscale value allocation, and properly sorts the colors that cannot be distinguished by brightness mapping.

[0169] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A digital photo frame display intelligent adjustment system based on image content recognition, characterized in that: include: An acquisition module, configured to acquire the current display image of the digital photo frame; an identification module, configured to perform image content identification on the currently displayed image to obtain first content information; The recognition module includes: a grayscale module for performing image preprocessing on the current display image of the digital photo frame to obtain a corresponding grayscale image; an image processing module for: obtaining a first grayscale value of each pixel in the gradient image and screening out the largest first grayscale value; performing a convolution operation on the first grayscale value of each pixel with a preset horizontal gradient operator to obtain a horizontal gradient value of each pixel, and obtaining a first absolute value of the horizontal gradient value of each pixel; screening out the largest first absolute value and the smallest first absolute value, and calculating a first difference; subtracting the first absolute value of each pixel from the smallest first absolute value to obtain a plurality of second differences; calculating the ratio of the plurality of second differences to the first difference to obtain a second grayscale value of each pixel; performing a convolution operation on the first grayscale value of each pixel with a preset vertical gradient operator to obtain a vertical gradient value of each pixel, and obtaining a second absolute value of the vertical gradient value of each pixel; screening out the largest second absolute value and the smallest second absolute value, and calculating a third difference; subtracting the second absolute value of each pixel from the smallest second absolute value to obtain a plurality of fourth difference values; calculating the ratio of the plurality of fourth difference values to the second difference value to obtain a third grayscale value of each pixel; performing weighted summation processing on the second grayscale value and the third grayscale value of each pixel to obtain a target grayscale value of each pixel; determining the pixel value of each pixel based on the target grayscale value of each pixel; and generating a gradient image based on the pixel value of each pixel; a model training module, configured to: establish a content recognition model; obtain a sample image set, input the sample image set into the content recognition model, and perform iterative training according to a preset loss function; during the training of the content recognition model, obtain the number of trained iterations of the content recognition model, and determine whether the trained iteration number is greater than a preset iteration number; if it is determined that the trained iteration number is greater than the preset iteration number, adjust the parameters of the preset loss function to obtain a target loss function; iteratively train the content recognition model according to the target loss function until the number of iterations is completed to obtain a trained content recognition model; a fifth control module, configured to input the gradient image into the trained content recognition model and output corresponding first content information; The first control module is configured to determine whether the first content information is consistent with the preset second content information; when it is determined that the first content information is inconsistent with the preset second content information, query a database based on the second content information to obtain a target image including the second content information, and control the digital photo frame to display the target image.

2. The digital photo frame display intelligent adjustment system based on image content recognition according to claim 1, characterized in that: Also includes: A working state detection module, used for detecting whether the digital photo frame is in working state; A human body detection module, used to detect whether there is a human body around the digital photo frame; The second control module is connected to the working state detection module and the human body detection module respectively, and is used to turn off the digital photo frame when it is determined that the digital photo frame is in working state and there is no human body around the digital photo frame.

3. The digital photo frame display intelligent adjustment system based on image content recognition according to claim 2, characterized in that: The working status detection module includes: A brightness signal acquisition module, configured to acquire a brightness signal of the digital photo frame; The first judgment module is configured to analyze the brightness signal to obtain a brightness value, determine whether the brightness value is greater than a preset brightness value, and determine that the digital photo frame is in a working state when it is determined that the brightness value is greater than the preset brightness value.

4. The digital photo frame display intelligent adjustment system based on image content recognition according to claim 2, characterized in that: The human body detection module includes: An infrared energy detection module, used to detect the infrared energy value of the environment surrounding the digital photo frame; A second judgment module is used to judge whether the infrared energy value has a sudden change; a third control module, configured to obtain an image of a mutation region when the second judgment module determines that a mutation has occurred in the infrared energy value, perform human body recognition on the image of the mutation region according to human body recognition technology, obtain a recognition result, and determine whether there is a human body around the digital photo frame according to the recognition result.

5. The digital photo frame display intelligent adjustment system based on image content recognition according to claim 4, characterized in that: Before performing human body recognition on the image of the mutation region according to the human body recognition technology, the method further includes performing noise reduction processing on the image of the mutation region.

6. The digital photo frame display intelligent adjustment system based on image content recognition according to claim 1, characterized in that: Also includes: An environmental image acquisition module, used to acquire an environmental image around the digital photo frame; Ambient color temperature acquisition module, used for: Performing image segmentation processing on the environment image to obtain a plurality of sub-environment images; Obtaining the RGB color value of each pixel in the sub-environment image, wherein the RGB color value includes an R channel value, a G channel value, and a B channel value; The R channel average value is calculated based on the R channel value of each pixel in the sub-environment image; The G channel average value is calculated based on the G channel value of each pixel in the sub-environment image; The B channel average value is calculated based on the B channel value of each pixel in the sub-environment image; Calculating a ratio of the R channel average value to the G channel average value as a first ratio; Calculating a ratio of the B channel average value to the G channel average value as a second ratio; Generate a coordinate point according to the first ratio and the second ratio, mark the coordinate point on a preset color temperature map to obtain a marked point, obtain a color temperature value corresponding to the marked point, and use the color temperature value as the color temperature value of the sub-environment image; Comparing the color temperature values of the plurality of sub-environment images with a preset threshold value, selecting the sub-environment images whose color temperature values are greater than or equal to the preset threshold value, and generating a first set; Filtering sub-environment images with color temperature values smaller than a preset threshold and generating a second set; Calculating a first color temperature average value according to the color temperature values of the sub-environment images included in the first set; Calculating a second color temperature average value according to the color temperature values of the sub-environment images included in the second set; Counting a first number of sub-environment images included in the first set; Counting a second number of sub-environment images included in the second set; calculating a sum of the first quantity and the second quantity; calculating a third ratio of the first quantity to the sum, and using the third ratio as a first weight coefficient; calculating a fourth ratio of the second quantity to the sum, and using the fourth ratio as a second weight coefficient; Calculate an ambient color temperature value according to the first color temperature average value, the second color temperature average value, the first weight coefficient, and the second weight coefficient; A current color temperature value acquisition module, used to obtain the current color temperature value of the digital photo frame; a fourth control module, configured to query a preset ambient color temperature value-target color temperature value table according to the ambient color temperature value, obtain a corresponding target color temperature value, calculate a difference between the current color temperature value and the target color temperature value, use the difference as a color temperature adjustment coefficient, and adjust the current color temperature value of the digital photo frame according to the color temperature adjustment coefficient.

7. The digital photo frame display intelligent adjustment system based on image content recognition according to claim 1, characterized in that: Also includes: A contrast acquisition module, configured to acquire the contrast of the digital photo frame; The sixth control module is configured to determine whether the contrast is within a preset contrast range, and adjust the contrast when it is determined that the contrast is not within the preset contrast range.

8. The digital photo frame display intelligent adjustment system based on image content recognition according to claim 4, characterized in that: The determining whether the infrared energy value has a sudden change includes: Obtain infrared energy values at two adjacent time points, calculate the infrared energy difference, determine whether the infrared energy difference is within a preset infrared energy difference range, and if it is determined that the infrared energy difference is not within the set infrared energy difference range, determine that the infrared energy value has undergone a mutation.

9. The digital photo frame display intelligent adjustment system based on image content recognition according to claim 4, characterized in that: The denoising process for the mutation region image includes inputting the mutation region image into a pre-trained denoising model and outputting a denoised mutation region image.

Images