Method and apparatus for determining adaptive brightness for images containing animals

Abstract

A method for determining adaptive brightness for images containing animals, performed using a computing device, is disclosed. [Solution] The method may include the steps of: acquiring a first image containing an animal; identifying a first region corresponding to the animal in the entire region of the first image; calculating a first brightness value in the first region; and calculating an adaptive brightness of the first image based on the first brightness value.

Description

Technical Field

[0001] The present disclosure relates to a technology for image processing, and relates to a method and apparatus for determining an adaptive luminance for an image including an animal.

Background Art

[0002] Generally, an automatic luminance correction algorithm of a camera determines an exposure time of a sensor and a sensitivity value (ISO value) of the sensor based on an average value of the luminance of the entire input image, so as to capture a photograph with appropriate luminance. When acquiring biometric information of a pet with a mobile phone, when photographing a puppy or a cat having a black coat color, a conventional automatic luminance correction algorithm determines that the average luminance is dark, and by excessively correcting the exposure time and the sensor sensitivity value, there is a problem of acquiring low-quality biometric information with severe shaking and noise.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] The present disclosure aims to solve the problems caused by the above-described background art, and provides a method and apparatus for determining an adaptive luminance for an image including an animal.

Means for Solving the Problems

[0005] To achieve the above objectives, a method for determining adaptive luminance for an image containing an animal is disclosed, which is performed on a computing device according to one embodiment of the present disclosure. This method may include the steps of: acquiring a first image containing an animal; identifying a first region corresponding to the animal in the entire region of the first image; calculating a first luminance value in the first region; and calculating adaptive luminance for the first image based on the first luminance value.

[0006] In one embodiment, after the step of calculating the first luminance value, the step of calculating the adaptive luminance of the first image may further include a step of calculating a second luminance value in a second region which is the region excluding the first region from the entire region, and the step of calculating the adaptive luminance of the first image may include a step of calculating the adaptive luminance of the first image based on the first luminance value and the second luminance value.

[0007] In one embodiment, the step of calculating the adaptive brightness of the image based on the first brightness value and the second brightness value may include, if the first brightness value is greater than the second brightness value, the step of calculating a first modified brightness value by applying a first weight to the first brightness value, the step of calculating a second modified brightness value by applying a second weight to the second brightness value, and the step of calculating the adaptive brightness of the image based on the first modified brightness value and the second modified brightness value.

[0008] In one embodiment, the first weight and the second weight may be predetermined according to the type, breed, and coat color of the animal.

[0009] In one embodiment, the step of calculating the adaptive brightness of the image based on the first corrected brightness value and the second corrected brightness value may include the step of determining the average value of the first corrected brightness value and the second corrected brightness value as the adaptive brightness of the image.

[0010] In one embodiment, the step of calculating the adaptive brightness of the image based on the first brightness value and the second brightness value may include, if the first brightness value is smaller than the second brightness value, the step of calculating a third modified brightness value by applying a third weight to the first brightness value, and the step of calculating the adaptive brightness of the image based on the third modified brightness value.

[0011] In one embodiment, the third weight may be determined based on information from the camera that captured the first image.

[0012] In one embodiment, the step of calculating the first luminance value in the first region may include the steps of dividing the first region into a predetermined number of sub-regions and determining the average value of the luminance values ​​obtained in each of the sub-regions as the first luminance value.

[0013] In one embodiment, the step of identifying the first region may include the step of identifying the first region corresponding to the animal in the entire region of the first image using a pre-trained artificial intelligence-based animal detection model.

[0014] In one embodiment, the animal detection model may be pre-trained using training data that includes training images containing at least one animal and regions of the at least one animal corresponding to the training images.

[0015] In one embodiment, the method may further include the steps of: generating a second image by applying the adaptive brightness to the first image; identifying a third region corresponding to the animal in the entire area of ​​the second image; and determining the third region as the biological information of the animal.

[0016] In one embodiment, the step of calculating the adaptive brightness of the first image may include the step of calculating the adaptive brightness of the first image based on at least one piece of information from the animal's species, breed, and fur color, and the first brightness value.

[0017] In one embodiment, a computing device for determining adaptive brightness for an image containing an animal is disclosed. The computing device may include one or more processors and a memory for storing instructions that can be executed by the one or more processors. The one or more processors may perform the following steps: acquire a first image containing an animal; identify a first region corresponding to the animal in the entire region of the first image; calculate a first brightness value in the first region; and calculate adaptive brightness for the first image based on the first brightness value. [Effects of the Invention]

[0018] This disclosure describes a method for acquiring biological information images of pets, which involves separating the animal region from the background region and adaptively determining the brightness of the image based on the brightness information of each region.

[0019] The effects obtained from this disclosure are not limited to those described above, and other effects not mentioned above will be clearly understood by a person with ordinary skill in the art to which this disclosure pertains from the following description. [Brief explanation of the drawing]

[0020] [Figure 1] This is an illustrative diagram of a computing device for determining adaptive brightness for an image containing an animal, according to one embodiment of the present disclosure. [Figure 2] This is a flowchart illustrating a method for determining adaptive brightness for an image containing an animal, according to one embodiment of the present disclosure. [Figure 3] This is a flowchart illustrating a method for determining adaptive brightness for an image containing an animal, according to one embodiment of the present disclosure. [Figure 4] This is a flowchart illustrating a method for determining adaptive brightness for an image containing an animal, according to one embodiment of the present disclosure. [Figure 5]A flowchart for explaining a method for determining an adaptive luminance for an image including an animal according to an embodiment of the present disclosure. [Figure 6] A flowchart for explaining a method for determining an adaptive luminance for an image including an animal according to an embodiment of the present disclosure. [Figure 7] A flowchart for explaining a method for determining an adaptive luminance for an image including an animal according to an embodiment of the present disclosure. [Figure 8] A diagram showing a process for determining an adaptive luminance for an image included in an animal according to an embodiment of the present disclosure.

Mode for Carrying Out the Invention

[0021] Various embodiments will be described with reference to the drawings. In this specification, various explanations are presented to provide an understanding of the present disclosure. However, it is clear that these embodiments can be implemented without such specific explanations.

[0022] The term "or" is intended to mean an inclusive "or" rather than an exclusive "or". That is, unless otherwise specified or unclear in context, "X uses A or B" is intended to mean one of the natural inclusive substitutions. That is, "X uses A or B" applies to any of these cases when X uses A, X uses B, or X uses both A and B. Further, the term "and / or" used in this specification should be understood to refer to all possible combinations of one or more of the listed related items and to include them.

[0023] Also, the terms "include" and / or "contain" should be understood to mean that the feature and / or component exists. However, the terms "include" and / or "contain" should be understood not to exclude the presence or addition of one or more other features, components, and / or groups thereof.

[0024] Furthermore, unless otherwise specified or it is not clear from the context that it refers to the singular form, the singular in this specification and claims should generally be interpreted as meaning "one or more."

[0025] In this disclosure, terms such as the first, second, or third, and so on, are used to distinguish between multiple entities. The entities represented as the first and second may be identical or different from one another.

[0026] The risks in this disclosure may be all risks that may arise from the replacement of human judgment by the model during the development and operation of the model-based service. In one embodiment, the risks may arise from existing business procedures, information protection, security, etc., or may be the remaining risks after excluding existing risks that already exist. In one embodiment, the risks may be determined based on the core value of the model to risk management, in light of industry-specific characteristics. For example, in the financial industry, the risks may be determined based on customer fundamental rights such as property rights, equality rights, and transparency. In one embodiment, the risks may be the expected residual risks that remain after controls have been established for the detailed risks identified in the model-based service.

[0027] In this disclosure, the terms "artificial intelligence-based model" (e.g., an animal detection model) can be used interchangeably. An artificial intelligence-based model can consist of a set of interconnected computational units, generally called nodes. These nodes are sometimes also called neurons. An artificial intelligence-based model consists of at least one node. The nodes (or neurons) that make up an artificial intelligence-based model can be interconnected by one or more links.

[0028] In an artificial intelligence-based model, one or more nodes connected via links can form relative input and output node relationships. The concepts of input and output nodes are relative; any node that is an output node to another node can be an input node to another node, and vice versa. As mentioned above, the relationship between input and output nodes can be generated around links. One input node can be connected to one or more output nodes via links, and vice versa.

[0029] In a relationship between input and output nodes connected via a single link, the data of the output node can be determined based on the data input to the input node. Here, the link interconnecting the input and output nodes may have weights. The weights may be variable and can be varied by the user or algorithm to enable the artificial intelligence-based model to perform the desired function. For example, if one or more input nodes are interconnected to one output node by their respective links, the output node's value can be determined based on the values ​​input to the input nodes connected to the output node, and the weights set for the links corresponding to each input node.

[0030] Artificial intelligence-based models can include convolutional neural networks (CNNs), recurrent neural networks (RNNs), autoencoders, Generative Adversarial Networks (GANs), and others.

[0031] Artificial intelligence-based models can be trained using at least one of the following methods: supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning. Training an artificial intelligence-based model can be a process of applying knowledge to the model to enable it to perform specific actions.

[0032] Artificial intelligence-based models can learn to minimize output errors. Model training involves repeatedly inputting training data into the model, calculating the model's output and target error for the training data, and updating the weights of each node in the model by backpropagating the model's error from the output layer to the input layer in a way that reduces the error. In supervised learning, training data with the correct answer labeled is used (i.e., labeled training data), while in unsupervised learning, the training data may not be labeled. For example, in supervised learning for data classification, the training data may be data with a category labeled for each data point. Labeled training data is input into an artificial intelligence-based model, and the error can be calculated by comparing the model's output (category) with the labels on the training data. As another example, in unsupervised learning for data classification, the error can be calculated by comparing the input training data with the output of the artificial intelligence-based model. The calculated errors are backpropagated in the AI-based model in the reverse direction (i.e., from the output layer to the input layer), and the connected weights of each node in each layer of the AI-based model can be updated according to the backpropagation. The amount of change in the connected weights of each node to be updated can be determined according to the learning rate. The calculation of the AI-based model on the input data and the backpropagation of errors can constitute a learning cycle (epoch). The learning rate can be applied differently depending on the number of iterations of the learning cycle of the AI-based model. For example, a high learning rate can be used in the early stages of learning the AI-based model to increase efficiency by quickly ensuring that the AI-based model achieves a certain level of performance, while a low learning rate can be used in the later stages of learning to improve accuracy.

[0033] Figure 1 is an illustrative diagram of a computing device for determining adaptive brightness for an image containing an animal, according to one embodiment of the present disclosure.

[0034] The computing device 100 may include one or more processors 110, memory 130, and network unit 150.

[0035] The processor 110 can consist of one or more cores. The processor 110 can control the overall operation of the computing device 100. The processor 110 can read a computer program stored in memory 130 and determine an adaptive brightness for an image containing an animal according to one embodiment of the present disclosure.

[0036] Memory 130 can store any form of information generated or determined by the processor 110, and any form of information received via the network unit 150. Memory 130 may include at least one type of storage medium from among flash memory type, hard disk type, multimedia card micro type, card type memory (SD, XD memory, etc.), RAM (Random Access Memory), SRAM (Static Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), PROM (Programmable Read-Only Memory), magnetic memory, magnetic disk, and optical disk.

[0037] The network unit 150 may include any wired or wireless network capable of sending and receiving data, information, and signals in any form. The network unit 150 can communicate with external devices. These external devices may be, for example, servers or user devices that provide risk assessments for artificial intelligence-based models.

[0038] In this disclosure, a computer (for example, a computing device 100) typically includes various computer-readable media. Any media accessible by a computer can be a computer-readable medium. Such computer-readable media include volatile and non-volatile media, transient and non-transitory media, and portable and non-portable media. As an example that is not limiting, computer-readable media may include computer-readable storage media. Computer-readable storage media include volatile and non-volatile media, transient and non-transitory media, portable and non-portable media implemented in any way or technique for storing information such as computer-readable instructions, data structures, program modules, or other data. Computer-readable storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, DVD (digital video disk) or other optical disc storage devices, magnetic cassettes, magnetic tapes, magnetic disk storage devices or other magnetic storage devices, or any other media accessible by a computer and usable for storing desired information.

[0039] Next, a specific process performed by the processor 110 to determine the adaptive brightness for an image containing an animal, according to one embodiment of the present disclosure, will be described below.

[0040] Figures 2 to 7 are flowcharts illustrating a method for determining adaptive luminance for an image containing an animal, according to one embodiment of the present disclosure. The steps shown in Figures 2 to 7 are exemplary. It will therefore be apparent to those skilled in the art that some of the steps in Figures 2 to 7 may be omitted, or additional steps may exist, within the scope of the spirit of the present disclosure. The flowcharts shown in Figures 2 to 7 may be performed, for example, by a computing device 100.

[0041] Referring to Figure 2, the processor 110 of the computing device 100 can acquire a first image containing an animal (S100). In one embodiment, an animal is a taxonomic group of organisms that is in contrast to plants, and can be a multicellular, eukaryotic organism. An animal can include not only pets, but also cats, dogs, birds, and the like.

[0042] For example, the processor 110 can acquire a first image containing an animal by photographing the animal through the imaging unit (e.g., a camera) of the computing device 100.

[0043] As another example, the processor 110 can acquire a first image containing an animal from images pre-stored in the memory 130.

[0044] As another example, the processor 110 can receive a first image containing an animal from an external device different from the computing device 100 (e.g., a server, a user terminal, etc.).

[0045] The processor 110 can identify the first region corresponding to the animal in the entire area of ​​the first image (S200).

[0046] In one embodiment, referring to Figure 3, the processor 110 can identify a first region corresponding to an animal in the entire region of the first image using a pre-trained artificial intelligence-based animal detection model (S210).

[0047] In one embodiment, the animal detection model can be pre-trained using training data that includes training images containing at least one animal and regions of at least one animal corresponding to the training images. In one embodiment, the processor 110 can detect the type and location of an animal using information about the animal output from the animal detection model to which the image has been input (e.g., location information of the animal, degree of identification and classification of the animal, etc.).

[0048] In some embodiments of the present disclosure, if information about an animal is mapped to the first image, the processor 110 can identify a first region corresponding to an animal in the entire region of the first image based on the information about the animal mapped to the first image, using a first animal detection model trained on a training dataset corresponding to the animal mapped to the first image among a plurality of animal detection models.

[0049] In some embodiments, the first animal detection model can be pre-trained using training data that includes a first training image containing animals mapped to a first image, and the region of the animal corresponding to the first training image. By using an animal detection model corresponding to each animal, the computing device 100 can achieve higher accuracy than when using a general-purpose animal detection model. Furthermore, by using different training data depending on the animal for each animal detection model, the computing device 100 can reduce the amount of training and training time compared to an animal detection model that trains for all animals.

[0050] Referring again to Figure 2, the processor 110 can calculate the first brightness value in the first region (S300).

[0051] Referring to Figure 4, the processor 110 can divide the first region into a predetermined number of sub-regions (S310). For example, the processor 110 can divide the first region into nine sub-regions.

[0052] In some embodiments of this disclosure, the processor 110 can divide the first region into a predetermined number of subregions depending on the type and breed of the animal. For example, if the animal is a dog and the breed is a stew, the processor 110 can divide it into a predetermined number of 10 subregions. The number of subregions to be divided may vary depending on the type and breed of the animal. For example, if the animal has different colors in different regions of its face, or if the body and face have different colors, the processor 110 can increase the number of subregions to be divided beyond a baseline number (e.g., 5), as this requires more detailed analysis. In another example, if the animal has similar colors in different regions, the processor 110 can reduce the number of subregions to be divided beyond a baseline number.

[0053] In some embodiments of this disclosure, if the classification of animals classified using an animal detection model is a multi-class classification (for example, if there is a 30% chance that it is a deer, a 30% chance that it is a horse, and a 40% chance that it is a roe deer), the first region can be divided into a criterion number of subregions.

[0054] The processor 110 can determine the average value of the brightness values ​​acquired in each of the multiple sub-regions as the first brightness value (S320). In one embodiment, the processor 110 can determine the lowest (or highest) brightness value among the brightness values ​​acquired in each of the multiple sub-regions as the first brightness value.

[0055] Referring again to Figure 2, the processor 110 can calculate the second luminance value in the second region, which is the region excluding the first region in the entire area (S400). For example, the processor 110 can divide the second region into a predetermined number of sub-regions. The processor 110 can determine the second luminance value as either the average, minimum, or maximum value of the luminance values ​​obtained in each of the sub-regions.

[0056] The processor 110 can calculate the adaptive brightness of the first image based on the first brightness value and the second brightness value (S500).

[0057] For example, referring to Figure 5, if the first brightness value is greater than the second brightness value, the processor 110 can calculate a first modified brightness value by applying a first weight to the first brightness value (S510).

[0058] The processor 110 can calculate a second modified brightness value by applying a second weight to the second brightness value (S520).

[0059] In one embodiment, the first and second weights can be predetermined according to the type, breed, and coat color of the animal. For example, if the animal's color is in the dark color family, the processor 110 may set the first weight higher than the second weight. In another example, if the animal's color is in the light color family, the processor 110 may set the first weight lower than the second weight.

[0060] The processor 110 can calculate the adaptive brightness of the image based on the first corrected brightness value and the second corrected brightness value (S530).

[0061] In one embodiment, the processor 110 can determine the average value of a first corrected brightness value and a second corrected brightness value as the adaptive brightness of the image.

[0062] As another example, the processor 110 may determine the adaptive brightness by considering only the areas where animals are present, without further consideration of areas where no animals are present. In one embodiment, the processor 110 can acquire a first image containing animals. The processor 110 can identify a first region corresponding to an animal within the entire area of ​​the first image. The processor 110 can calculate the adaptive brightness of the first image based on the first brightness value of the first region.

[0063] Specifically, the processor 110 can calculate the adaptive brightness of the first image based on one or more pieces of information from the animal's species, breed, and coat color, and a first brightness value. For example, the processor 110 can calculate the adaptive brightness of the first image by multiplying the first brightness value by a coefficient predetermined based on one or more pieces of information from the animal's species, breed, and coat color (e.g., 1.1, 0.9, etc.). As another example, the processor 110 can calculate the adaptive brightness of the first image by multiplying the first brightness value by a coefficient predetermined for a combination of animal species, breed, and coat color (e.g., 1.1, 0.9, etc.). For example, the predetermined coefficient for the combination (dog, Doberman, black) might be 0.9. As yet another example, the predetermined coefficient for the combination (cat, Persian, white) might be 1.1.

[0064] The processor 110 can divide the first region into a predetermined number of sub-regions. The processor 110 can determine the average value of the brightness values ​​obtained in each of the sub-regions as the first brightness value.

[0065] Specifically, referring to Figure 6, if the first luminance value is smaller than the second luminance value, the processor 110 can calculate a third modified luminance value by applying a third weight to the first luminance value (S540). The fact that the first luminance value is smaller than the second luminance value can be interpreted as meaning that the first region corresponding to the animal is dark, and therefore the luminance value needs to be adjusted based on the first region.

[0066] In one embodiment, the third weight can be determined based on information from the camera that captured the first image. In one embodiment, the camera information may include lens information, sensor information (e.g., sensor exposure time, sensitivity value, etc.), and so on.

[0067] The processor 110 can calculate the adaptive brightness of the image based on the third corrected brightness value (S550).

[0068] In one embodiment, referring to Figure 7, the processor 110 can generate a second image by applying adaptive brightness to the first image (S600). For example, the processor 110 can generate the second image by converting the brightness of the first image to adaptive brightness.

[0069] The processor 110 can identify a third region corresponding to an animal in the entire area of ​​the second image (S700). For example, the processor 110 can identify a third region corresponding to an animal in the entire area of ​​the second image using a pre-trained artificial intelligence-based animal detection model.

[0070] The processor 110 can determine the third domain as the animal's biological information (S800). For example, the processor 110 can determine the structure and color of the animal's eyes, the color and pattern of its fur, its body shape, its size, etc., which are located in the third domain, as the animal's biological information.

[0071] Figure 8 shows the process for determining adaptive brightness for an image contained in an animal according to one embodiment of the present disclosure. Referring to Figure 8, some of the configurations described later that have already been explained can be omitted. Specific explanations of the configurations described later with reference to Figure 8 can be replaced with the explanations given above with reference to Figures 1 to 7.

[0072] Referring to Figure 8, the processor 110 can acquire the first image 210 which contains the animal.

[0073] The processor 110 can identify a first region 220 corresponding to an animal within the entire region of the first image 210. For example, the processor 110 can identify a first region 220 corresponding to an animal within the entire region of the first image using a pre-trained artificial intelligence-based animal detection model.

[0074] The processor 110 can calculate a first brightness value 240 in the first region 220.

[0075] The processor 110 can calculate a second brightness value 250 in the second region 230, which is the remaining region of the first image 210 excluding the first region 220.

[0076] The processor 110 can calculate the adaptive brightness 260 of the first image 210 based on the first brightness value 240 and the second brightness value 250.

[0077] The description of the embodiments presented is provided so that a person with ordinary skill in the art of this disclosure may utilize or practice this disclosure. Various modifications of these embodiments will be obvious to a person with ordinary skill in the art of this disclosure. The general principles defined herein can be applied to other embodiments without departing from the scope of this disclosure. Thus, this disclosure is not limited to the embodiments presented herein. This disclosure should be interpreted in the broadest sense, consistent with the principles and novel features presented herein.

Claims

1. A method for determining adaptive brightness for an image containing an animal, which is performed on a computing device, The steps include obtaining a first image containing an animal, The steps include identifying a first region corresponding to the animal in the entire region of the first image, A step of calculating a first luminance value in the first region, A method comprising the step of calculating the adaptive brightness of a first image based on the first brightness value.

2. After the step of calculating the first luminance value, The method further includes the step of calculating a second luminance value in the second region, which is the region excluding the first region from the entire region, The step of calculating the adaptive brightness of the first image is: The method according to claim 1, further comprising the step of calculating an adaptive brightness of the first image based on the first brightness value and the second brightness value.

3. The step of calculating the adaptive brightness of the image based on the first brightness value and the second brightness value is: If the first luminance value is greater than the second luminance value, the first modified luminance value is calculated by applying a first weight to the first luminance value. The steps include: calculating a second modified luminance value by applying a second weight to the second luminance value; The method according to claim 2, comprising the step of calculating an adaptive brightness of the image based on the first corrected brightness value and the second corrected brightness value.

4. The method according to claim 3, wherein the first weight and the second weight are predetermined according to the type, breed, and coat color of the animal.

5. The step of calculating the adaptive brightness of the image based on the first corrected brightness value and the second corrected brightness value is: The method according to claim 3, further comprising the step of determining the average value of the first corrected brightness value and the second corrected brightness value as the adaptive brightness of the image.

6. The step of calculating the adaptive brightness of the image based on the first brightness value and the second brightness value is: If the first luminance value is smaller than the second luminance value, the third modified luminance value is calculated by applying a third weight to the first luminance value. The method according to claim 3, comprising the step of calculating the adaptive brightness of the image based on the third corrected brightness value.

7. The method according to claim 6, wherein the third weight is determined based on information from the camera that captured the first image.

8. The step of calculating the first luminance value in the first region is: The steps include dividing the first region into a predetermined number of sub-regions, The method according to claim 1, comprising the step of determining the average value of the luminance values ​​obtained in each of the plurality of sub-regions as the first luminance value.

9. The step of identifying the first region is: The method according to claim 1, comprising the step of identifying a first region corresponding to the animal in the entire region of the first image using a pre-trained artificial intelligence-based animal detection model.

10. The aforementioned animal detection model is The method according to claim 9, wherein the model is pre-trained using training data comprising training images containing at least one animal and regions of the at least one animal corresponding to the training images.

11. The steps include generating a second image by applying the adaptive brightness to the first image, The steps include identifying a third region corresponding to the animal in the entire region of the second image, The method according to claim 1, further comprising the step of determining the third region as biological information of the animal.

12. The step of calculating the adaptive brightness of the first image is: The method according to claim 1, comprising the step of calculating adaptive brightness of a first image based on at least one piece of information from the type, breed, and coat color of the animal and the first brightness value.

13. A computing device for determining adaptive brightness for images containing animals, One or more processors, Includes a memory for storing instructions that can be executed by one or more processors, The one or more processors described above are: The process of obtaining the first image containing an animal, A process of identifying a first region corresponding to the animal in the entire region of the first image, The process of calculating the first luminance value in the first region, A computing device that performs the process of calculating the adaptive brightness of the first image based on the first brightness value.

Images