Artificial intelligence method for detecting gonad maturity in mussels

Abstract

The present invention relates to a method, to a computer program product and to a system for detecting gonad maturity in adult mussels. The method comprises the following steps: a) collecting adult mussels from an area of interest; b) extracting fresh gonad tissue samples; c) preparing smears; d) capturing images of the smears with a portable microscope and a digital camera; e) analysing the images using an application on a smart device, which transmits said images to a server equipped with a convolutional neural network model, in order to classify the samples as female, male or undetermined, and predict the degree of maturity in the females; and f) receiving, on the smart device, the results of the sex classification and the degree of reproductive maturity in the analysed samples. The invention further discloses a method for capturing mussel spat that comprises detecting gonad maturity in adult mussels.

Description

[0001] “Artificial intelligence method for detecting gonadal maturity in adult mussels and collecting mussel seed”

[0002] DESCRIPTIVE MEMORANDUM

[0003] SCOPE

[0004] The present invention relates to the aquaculture industry, specifically mussel farming, focusing on the optimization of processes for the collection of mussel seeds.

[0005] BACKGROUND

[0006] Uncertainty and variability in mussel seed collection represent a critical global challenge for the industry, and are a key factor in ensuring the sustainability and productivity of mussel farming. Mussel reproductive cycles, influenced by bio-oceanographic factors such as temperature, salinity, and phytoplankton availability, are highly variable and difficult to predict. This unpredictability leads to inefficiencies in collector installation, with significant impacts on seed production and supply.

[0007] Anomalous reproductive events, such as spawning outside of typical periods or fluctuations in larval abundance, highlight the complexity of efficiently managing collection processes. However, traditional analytical methods rely on laborious processes that are difficult to scale up and have limitations in terms of accessibility and reproducibility.

[0008] The traditional method for assessing the reproductive status of mussels involves the extraction and visual analysis of their gonads using microscopic observation. This process begins with the collection of mussels in the field, followed by their preparation in the laboratory by opening and smearing them to obtain gonadal tissue samples. The samples are placed on slides and observed under a microscope to identify morphological characteristics that distinguish between males and females, as well as to determine the degree of gonadal maturity. This analysis, highly dependent on the technician's experience, allows for the identification of the periods when the gonads reach their maximum maturity, indicating the most suitable time for spawning and larval collection.However, this methodology has significant limitations, such as its dependence on trained but scarce personnel with limited availability, in addition to the time required to process each sample, which reduces the responsiveness to make timely decisions in the installation of collectors.

[0009] Preliminary art

[0010] Addressing this issue, the publication by Perez, Daniel & Lorenzo, Silvia & Alves, Dario & Fuentes, José & González-Fernández, Africa (2007) is noteworthy. “Adaptation and optimization of immunodetection by indirect immunofluorescence of mussel larvae using monoclonal antibodies.” Immunodetection is a more advanced technique that uses specific antibodies to identify mussel larvae, allowing for the precise differentiation of mussel larvae from other species. However, it requires specialized equipment, as well as laboratories equipped with fluorescence microscopes and personnel trained in immunological techniques. This entails high costs due to the reagents and equipment, limiting its use to well-funded laboratories. Furthermore, the need for specialized facilities prevents its direct application in remote or real-time environments.

[0011] Also known is the international patent WO2016207857A1, which describes a set of primers and a method for the detection and identification of mussel species of the genus Mytilus using high-resolution melting and PCR techniques. This method is based on the detection of polymorphisms in the polyphenolic adhesive protein gene, allowing differentiation between species such as Mytilus galloprovincialis, Mytilus chilensis, and Mytilus edulis. While this method allows for the precise identification of species and populations and provides data on genetic diversity and population structure, its implementation requires expensive equipment and reagents, as well as specialized personnel with knowledge of molecular genetics. Furthermore, it is unsuitable for field use, as it also requires a laboratory setting.

[0012] Technical problem addressed by the invention

[0013] Although multiple methods exist for detecting and identifying mussel larvae, they have limitations in terms of accuracy, cost, accessibility, and applicability in different environments. Reliance on specialized infrastructure, expert personnel, and laborious processes reduces the effectiveness of these methods, especially in remote or resource-limited regions. These limitations highlight the need for solutions that overcome current challenges and enable more efficient and sustainable seed collection globally, as well as a simple and effective method for determining the optimal time for mussel seed collection.

[0014] Technical Solution of the present invention

[0015] An object of the present invention is a method for detecting the gonadal maturity of adult mussels.

[0016] Another object of the present invention is a computer program product for detecting the gonadal maturity of adult mussels.

[0017] Another additional object of the present invention is a system for detecting the gonadal maturity of adult mussels.

[0018] Yet another additional object of the present invention is a method for collecting mussel seeds that includes detecting the gonadal maturity of adult mussels.

[0019] Advantages of the present invention

[0020] The present invention allows for the precise and rapid determination of the maturity stage of mussels. The present invention enables remote, efficient, and scalable access thanks to its portable and mobile design.

[0021] The present invention allows mussel farmers to accurately anticipate the ideal times to install collectors, maximizing seed collection and optimizing resources.

[0022] The present invention allows for decision-making in the mussel farming industry regarding the placement of collectors for seed collection, optimizing the seed collection stage through periodic sampling that provides indicators of when larvae may be available in the water.

[0023] BRIEF DESCRIPTION OF THE FIGURES

[0024] The accompanying figures are briefly described below to exemplify and illustrate the present invention. It should be noted that the figures are provided for illustrative purposes only and that the invention is not limited by these illustrations.

[0025] Figure 1 illustrates a microscope photograph of a mature female gonad smear from a mussel.

[0026] Figure 2 illustrates a microscope photograph of the male gonad smear of a mussel.

[0027] Figure 3 illustrates a microscope photograph of an indeterminate gonad smear from a mussel.

[0028] Figure 4 illustrates a microscope photograph of a female gonad smear during spawning from a mussel.

[0029] DETAILED DESCRIPTION OF THE INVENTION

[0030] The present invention is described below in detail, with reference to the accompanying figures. Well-known components, materials, or methods are not necessarily described in great detail to avoid obscuring the present description. Any specific structural and functional details described herein should not be construed as limiting, but merely as a basis for the claims and as a representative basis for teaching a person skilled in the art how to employ the invention in various ways.

[0031] According to the present invention, a method is provided for detecting the gonadal maturity of adult mussels comprising the steps of: a) collecting adult mussels in an area of ​​interest; b) extracting fresh samples of gonadal tissue from said adult mussels; c) preparing smears with said fresh samples; d) taking photographs of said fresh smears with a portable or pocket microscope and a digital camera; e) running, with a phone or smart device loaded with said photographs in its memory, an application configured to connect to and transmit said photographs to a server or server service, wherein said server or server service comprises an application configured to: i.analyze said photographs with a convolutional neural network model, previously trained to classify gonadal smear images of mussels as female, male and indeterminate, and predict the degree of reproductive maturity of female samples from smear images classified as female; and.

[0032] i. communicate to a telephone or mobile device, the determination of the sex as female, male or indeterminate of said adult mussels and the degree of reproductive maturity of samples determined as female, from the analysis and classification of said photographs taken of said fresh samples; f) receipt on said telephone or smart device, of a response from the server or server service with the determination of the sex of said adult mussels and the degree of maturity of the female samples.

[0033] According to the present invention, said convolutional neural network model is also pre-trained to classify as an unwanted image photographs taken in stage d) in a mistaken or erroneous manner and which do not correspond to a fresh smear sample of fresh samples.

[0034] According to the present invention, a computer program product is provided for detecting the gonadal maturity of adult mussels, configured to: i. analyze photographs of fresh smears of fresh samples of adult mussels with a convolutional neural network model, previously trained to classify images of mussel gonadal smears as female, male, and indeterminate, and predict the degree of reproductive maturity of female samples from images of smears classified as female; and

[0035] i. communicate to a telephone or mobile device, the determination of the sex as female, male or indeterminate of said adult mussels, the degree of reproductive maturity of samples determined as female, from the analysis and classification of said photographs.

[0036] According to the present invention, said computer program product is also configured with an API to utilize said convolutional neural network model.

[0037] According to the present invention, said computer program product is also configured to connect to and receive photographs from a phone or mobile device.

[0038] Additionally, according to the invention, said convolutional neural network model is also pre-trained to classify as unwanted images photographs taken in stage d) in a mistaken or erroneous manner and which do not correspond to a fresh smear sample of fresh samples.

[0039] According to the present invention, a system for detecting the gonadal maturity of adult mussels is provided, comprising: a portable or pocket microscope; a digital camera; and a phone or smart device loaded with an application configured to connect to and transmit photographs to a server or server service, wherein said server or server service comprises an application configured: i. to analyze photographs of fresh smears of fresh samples of adult mussels with a convolutional neural network model, previously trained to classify images of mussel gonadal smears as female, male, and indeterminate, and to predict the degree of reproductive maturity of female samples from images of smears classified as female; and ii.to communicate to a telephone or mobile device, the determination of the sex as female, male or indeterminate of said adult mussels, the degree of reproductive maturity of samples determined as female, from the analysis and classification of said photographs.

[0040] According to one form of the invention, said digital camera is integrated into said phone or smart device which includes an application for detecting the gonadal maturity of adult mussels.

[0041] According to another form of the invention, said digital camera is integrated into said portable microscope which includes wired or wireless communication means with a phone or smart device.

[0042] A smartphone or smart device is defined as any portable electronic device that integrates data processing and storage capabilities, as well as wireless connectivity, such as smartphones, tablets, or similar devices, enabling the capture, storage, processing, and transmission of information. These smartphones or smart devices may include sensors, integrated cameras, and access to applications.According to the present invention, a method for collecting mussel seeds is provided, which includes detecting the gonadal maturity of adult mussels in the following steps: a) collecting adult mussels in an area of ​​interest; b) extracting fresh samples of gonadal tissue from said adult mussels; c) preparing smears with said fresh samples; d) taking photographs of said fresh smears with a portable microscope and a digital camera; e) running, with a phone or smart device loaded with said photographs in its memory, an application configured to connect to and transmit said photographs to a server or server service, wherein said server or server service comprises an application configured to: i.analyze said photographs with a convolutional neural network model, previously trained to classify gonadal smear images of mussels as female, male and indeterminate, and predict the degree of reproductive maturity of female samples from smear images classified as female; and.

[0043] (i) communicating to a telephone or mobile device the determination of the sex as female, male, or indeterminate of said adult mussels and the degree of reproductive maturity of samples determined to be female, based on the analysis and classification of said photographs taken of said fresh samples; (f) receiving on said telephone or smart device a response from the server or server service with the determination of the sex of said adult mussels and the degree of maturity of the female samples; (g) installing collectors in the area of ​​interest if the degree of maturity is estimated to be sufficient, during the optimal period to maximize the capture of larvae; (h) collecting the collectors with the mussel spat after a settling period, for example, 10 to 20 days.

[0044] Mussel collectors are structures designed to capture mussel larvae during their settlement stage. They are generally constructed from materials such as nets, ropes, or plastic mesh, providing a suitable surface for the larvae to attach and develop.

[0045] Example of training a convolutional neural network model

[0046] According to the present invention, said convolutional neural network model is pre-trained to classify mussel gonadal smear images as female, male and indeterminate, and to predict the maximum gonadal maturity state of mussels, with an image bank formed during a determined period of intensive sampling, for example, one year, collecting data at different predetermined sites.

[0047] For example, 43 bi-weekly samplings were conducted (20 in a natural area and 23 in a breeding center), collecting approximately 100 individuals per sample. Each individual was measured, weighed, sexed, and photographed. This resulted in a robust database with over 5,000 images of fresh smears captured using various devices, such as a laboratory microscope with a Motic camera and a portable microscope with 175X magnification connected to an Android or iPhone smartphone, such as the Carson MicroFlip 100X-250X pocket microscope.

[0048] Each image was labeled by experts to classify it as "male," "female," "undetermined," or "spawning." In addition, in the case of females, the degree of gonadal maturity was recorded, providing detailed labeling and a basis for training the model.

[0049] The images were standardized and segmented to ensure consistency in the input data. Data augmentation techniques, such as rotation, scaling, and brightness adjustments, were implemented to diversify the samples and strengthen the model against variations.

[0050] As an example of images used for the present invention, Figure 1 shows a photograph of fresh smears of mature female gonad, Figure 2 shows a photograph of fresh smears of male gonad, Figure 3 shows a photograph of fresh smears of indeterminate gonad, and Figure 4 shows a photograph of fresh smears of spawning gonad, all gonads of the species Mytilus chilensis.

[0051] The model was evaluated using a validation dataset to measure its accuracy, achieving prediction levels above 90% in the classification of mature females.

[0052] Although embodiments of the invention have been illustrated and described in detail in the drawings and description above, such illustration and description should be considered illustrative or exemplary and not restrictive; the invention is not limited to the examples particularly described. Persons skilled in the art understand and can make other variations of the examples particularly described from a study of the drawings, description, and claims. In the specification and claims, the terms "comprising" and "including" are open-ended and do not exclude other elements or steps, and the indefinite article "a" does not exclude a plurality.

Claims

CLAIMS 1.- Method for detecting gonadal maturity in adult chonta palms, CHARACTERIZED in that it comprises the steps of a) collecting adult chonta palms in an area of ​​interest; b) extracting fresh gonadal tissue samples from said adult chonta palms; c) preparing smears with said fresh samples; d) taking photographs of said fresh smears with a portable or pocket microscope and a digital camera; e) running, with a phone or smart device loaded with said photographs in its memory, an application configured to connect to and transmit said photographs to a server or server service, wherein said server or server service comprises an application configured to: i. analyze said photographs with a convolutional neural network model, previously trained to classify gonadal smear images of chonta palms as female, male, and indeterminate, and predict the degree of reproductive maturity of female samples from smear images classified as female; and i. communicate to a telephone or mobile device, the determination of the sex as female, male or indeterminate of said adult chontas and the degree of reproductive maturity of samples determined as female, from the analysis and classification of said photographs taken of said fresh samples; f) receipt on said telephone or smart device, of a response from the server or server service with the determination of the sex of said adult chontas and the degree of maturity of the female samples. 2.- Method according to claim 1, CHARACTERIZED in that the convolutional neural network model is also previously trained to classify as an unwanted image photographs taken in step d) in a mistaken or erroneous manner and which do not correspond to a fresh smear sample of fresh samples. 3.- Product of a computer program to detect the gonadal maturity of adult chonta palms, CHARACTERIZED in that it is configured to: i. analyze photographs of fresh smears of fresh samples of adult chonta palms with a convolutional neural network model, previously trained to classify images of gonadal smears of chonta palms as female, male and indeterminate, and predict the degree of reproductive maturity of female samples from images of smears classified as female; and i. communicate to a telephone or mobile device, the determination of the sex as female, male or indeterminate of said adult chontas, the degree of reproductive maturity of samples determined as female, from the analysis and classification of said photographs.

4. The computer program product according to claim 3, CHARACTERIZED in that it is also configured with an API to utilize said convolutional neural network model.

5. The computer program product according to claim 3, CHARACTERIZED in that said computer program product is also configured to connect to and receive photographs from a phone or mobile device.

6. The computer program product according to claim 3, CHARACTERIZED in that said convolutional neural network model is also previously trained to classify as an unwanted image photographs taken incorrectly and that do not correspond to a fresh smear sample of fresh samples.

7. System for detecting gonadal maturity in adult chonta palms, CHARACTERIZED in that it comprises: a portable or pocket microscope; a digital camera; and a phone or smart device loaded with an application configured to connect to and transmit photographs to a server or server service, wherein said server or server service comprises an application configured: i. to analyze photographs of fresh smears of fresh samples of adult mussels with a convolutional neural network model, previously trained to classify images of gonadal smears of mussels as female, male, and indeterminate, and to predict the degree of reproductive maturity of female samples from images of smears classified as female; and i. communicate to a telephone or mobile device, the determination of the sex as female, male or indeterminate of said adult mussels, the degree of reproductive maturity of samples determined as female, from the analysis and classification of said photographs.

8. The system according to claim 7, CHARACTERIZED in that said digital camera is integrated into said phone or smart device which includes an application for detecting the gonadal maturity of adult mussels.

9. The system according to claim 7, CHARACTERIZED in that said digital camera is integrated into said portable microscope which includes wired or wireless communication means with a telephone or smart device.

10. A method for collecting mussel seed that includes detecting the gonadal maturity of adult mussels, CHARACTERIZED in that it comprises the following steps: a) collecting adult mussels in an area of ​​interest; b) extracting fresh samples of gonadal tissue from said adult mussels; c) preparing smears with said fresh samples; d) taking photographs of said fresh smears with a portable microscope and a digital camera; e) running, with a phone or smart device loaded with said photographs in its memory, an application configured to connect to and transmit said photographs to a server or server service, where said server or server service comprises an application configured to: i. analyze said photographs with a convolutional neural network model, previously trained to classify mussel gonadal smear images as female, male and indeterminate, and predict the degree of reproductive maturity of female samples from smear images classified as female; and i. To communicate to a telephone or mobile device the determination of the sex as female, male, or indeterminate of said adult mussels and the degree of reproductive maturity of samples determined to be female, based on the analysis and classification of said photographs taken of said fresh samples; f) To receive on said telephone or smart device a response from the server or server service with the determination of the sex of said adult mussels and the degree of maturity of the female samples; g) To install collectors in the area of ​​interest if the degree of maturity is estimated to be sufficient, during the optimal period to maximize the capture of larvae; and h) To collect the collectors with the mussel spat after a period of settlement. 11.- Method according to claim 10, CHARACTERIZED in that the convolutional neural network model is also previously trained to classify as an unwanted image photographs taken in step d) in a mistaken or erroneous manner and which do not correspond to a fresh smear sample of fresh samples.

Images