Bacillus subtillis colony morphology identification and counting method based on deep learning

Abstract

The invention discloses a bacillus subtillis colony morphology identification and counting method based on deep learning, and relates to the technical field of microbial colonies, the method comprises the following steps: obtaining a colony image of to-be-identified bacillus subtillis; inputting the bacterial colony image into a trained bacterial colony detection model; outputting a form recognition result of the bacterial colony image by the bacterial colony detection model; counting bounding boxes of all bacterial colonies in the form recognition result to obtain a bacterial colony counting result and outputting the bacterial colony counting result; the bacterial colony detection model is based on a deep learning target detection model and is obtained by training a data set of bacillus subtillis bacterial colony images of various morphological categories. According to the method, the bacterial colony detection is realized based on the trained model through the image data set specially constructed by the polymorphic features of the bacillus subtillis, so that the technical problems that a traditional bacterial colony recognition technology depends on manpower and is low in efficiency are solved, accurate and automatic recognition and counting of bacterial colony morphology are realized, and the detection efficiency is improved. The efficiency and the reliability of microbial phenotype analysis are improved.

Description

Technical Field

[0001] This application relates to the field of microbial colony technology, and in particular to a method for identifying and counting Bacillus subtilis colony morphology based on deep learning. Background Technology

[0002] Microbial morphological classification and detection are fundamental steps in microbiological research and industrial applications. While computer image processing-based auxiliary analysis methods have automated the analysis process to some extent, they struggle to effectively segment and extract target regions when faced with real-world images exhibiting low contrast, uneven lighting, complex backgrounds, and overlapping cell structures. Furthermore, they are ill-suited to the highly variable and complex morphology of microorganisms.

[0003] With the development of artificial intelligence, especially deep learning technology, object detection and image classification models based on convolutional neural networks have provided a new technical path for the morphological classification of microorganisms. However, in complex scenarios such as the diverse morphology, dense distribution, and small size of Bacillus subtilis colonies, existing methods still struggle to achieve high-precision and high-efficiency automatic identification and counting.

[0004] There is a lack of effective methods for identifying and counting Bacillus subtilis colony morphology in the existing technology. Summary of the Invention

[0005] The purpose of this application is to provide a deep learning-based method for identifying and counting Bacillus subtilis colony morphology. This method utilizes an image dataset specifically designed for the multimorphic features of Bacillus subtilis, which is then used to train a YOLOv11 deep learning model. The trained model is then used to detect Bacillus subtilis colonies. This method solves the technical problems of traditional colony identification techniques, such as reliance on manual labor, low efficiency, and insufficient accuracy and robustness under complex imaging conditions. It achieves accurate and automated identification and counting of colony morphology, improving the efficiency and reliability of microbial phenotypic analysis.

[0006] To achieve the above objectives, this application provides the following solution: In a first aspect, this application provides a deep learning-based method for identifying and counting Bacillus subtilis colony morphology, comprising: acquiring a colony image of Bacillus subtilis to be identified; inputting the Bacillus subtilis colony image into a trained colony detection model; outputting the morphology recognition result of the colony image by the colony detection model, the morphology recognition result including the colony's bounding box, morphology classification, and confidence score; counting the bounding boxes of all colonies in the morphology recognition result to obtain and output the colony count result; wherein, the colony detection model is based on a deep learning target detection model and is trained using a dataset of Bacillus subtilis colony images of various morphological categories.

[0007] Optionally, the colony detection model is a YOLO series model.

[0008] Optionally, the YOLO series model is the YOLOv11 model.

[0009] Optionally, the various morphological categories include mucous-like, smooth without wrinkles or protrusions, colonies with a ring shape in the middle, and colonies with wrinkles or protrusions throughout.

[0010] Secondly, this application provides a training method for a deep learning-based Bacillus subtilis colony detection model, used to train the aforementioned colony detection model, comprising: acquiring colony images of Bacillus subtilis with multiple morphological categories; preprocessing and data augmenting the colony images; constructing a labeled dataset based on the preprocessed and data-augmented colony images; and training the colony detection model using the labeled dataset to obtain the trained colony detection model.

[0011] Optionally, the colony detection model is trained using transfer learning, and the colony detection model is fine-tuned using the weights of the pre-trained model.

[0012] Optionally, the preprocessing of the colony image includes image denoising, illumination correction, background removal, and extraction of colony regions using image segmentation algorithms.

[0013] Optionally, the methods for data augmentation of the colony images include: image rotation, flipping, brightness adjustment, and image size standardization.

[0014] Thirdly, this application provides a deep learning-based Bacillus subtilis colony morphology recognition and counting device, comprising: an acquisition module configured to acquire a colony image of Bacillus subtilis to be identified; an input module configured to input the Bacillus subtilis colony image into a trained colony detection model; a morphology result output module configured to output the morphology recognition result of the colony image by the colony detection model, the morphology recognition result including the colony bounding box, morphology classification, and confidence score; and a counting result output module configured to count the bounding boxes of all colonies in the morphology recognition result to obtain and output the colony counting result.

[0015] Thirdly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the deep learning-based Bacillus subtilis colony morphology recognition and counting method described above.

[0016] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application provides a deep learning-based method for identifying and counting Bacillus subtilis colony morphology. By using the colony detection model described above, it solves the problems of strong subjectivity, poor repeatability, and low efficiency in traditional manual identification and counting, achieving full automation and intelligence of the process. It can complete the analysis of colony images within seconds and output objective and consistent morphological classification and counting results, significantly improving detection efficiency and the reliability of the results. It adopts a deep learning-based target detection model as the core recognition engine and uses a dataset of Bacillus subtilis colony images including multiple morphological categories for specialized training. This solves the technical problem that traditional manual methods and general image processing algorithms cannot effectively learn and identify the complex and diverse specific morphological features of Bacillus subtilis, achieving accurate and automatic identification and classification of multiple colony morphologies. Its recognition accuracy is significantly better than that of traditional methods. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is an application environment diagram of a deep learning-based Bacillus subtilis colony morphology recognition and counting method in one embodiment of this application; Figure 2 A flowchart illustrating a deep learning-based method for identifying and counting Bacillus subtilis colony morphology, provided as an embodiment of this application; Figure 3 This is a comparison chart of the colony recognition effects of expert manual annotation and colony detection model in one embodiment of this application; Figure 4 A flowchart illustrating a training method for a deep learning-based Bacillus subtilis colony detection model provided in an embodiment of this application; Figure 5 This is a schematic diagram of four morphological types of bacterial colonies in one embodiment of this application; Figure 6 This is a schematic diagram illustrating the annotation of colony images in a model training dataset according to one embodiment of this application; Figure 7 This is a performance comparison chart of different models in one embodiment of this application; Figure 8 A schematic diagram of the functional modules of a Bacillus subtilis colony morphology recognition and counting device based on deep learning provided in an embodiment of this application; Figure 9This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0019] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0021] The deep learning-based Bacillus subtilis colony morphology recognition and counting method provided in this application can be applied to, for example... Figure 1 In the application environment shown, terminal 101 communicates with server 102 via a network. A data storage system can store the data that server 102 needs to process. The data storage system can be set up independently, integrated into server 102, or placed in the cloud or on another server. Terminal 101 can send colony images of Bacillus subtilis to be identified to server 102. Server 102 receives the colony images and inputs them into a trained colony detection model; it outputs the morphological recognition results of the colony images obtained by the colony detection model, including the colony bounding boxes, morphological classification, and confidence score; it then counts the bounding boxes of all colonies in the morphological recognition results to obtain and output the colony count results. Server 102 can feed back the obtained morphological recognition results and colony count results to terminal 101. In addition, in some embodiments, the morphological recognition results and colony counting results can also be implemented separately by the server 102 or the terminal 101. For example, the terminal 101 can directly target the colony image of Bacillus subtilis to be processed, or the server 102 can obtain the colony image of Bacillus subtilis to be identified from the data storage system.

[0022] The terminal 101 can be, but is not limited to, various desktop computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, and smart in-vehicle devices. Portable wearable devices can include smartwatches, smart bracelets, and head-mounted devices. The server 102 can be implemented using a standalone server or a server cluster composed of multiple servers, or it can be a cloud server.

[0023] In one exemplary embodiment, such as Figure 2 As shown, a deep learning-based method for identifying and counting Bacillus subtilis colony morphology is provided. This method is executed by a computer device, specifically a terminal or server, or both. In this embodiment, the method is applied to... Figure 1 We will use server 102 as an example to illustrate this.

[0024] like Figure 2 As shown, a deep learning-based method for identifying and counting Bacillus subtilis colony morphology includes the following steps: S11, Obtain colony images of the Bacillus subtilis to be identified; S12, Input the colony image of Bacillus subtilis into the trained colony detection model; S13, outputs the morphological recognition results of the colony image by the colony detection model. The morphological recognition results include the colony bounding box, morphological classification and confidence score. S14: Statistically analyze the bounding boxes of all colonies in the morphological recognition results, obtain the colony count results, and output them. The colony detection model is based on a deep learning-based target detection model and was trained using a dataset of Bacillus subtilis colony images of various morphological categories.

[0025] By implementing steps 11 to 14 above, and based on the colony detection model, the problems of strong subjectivity, poor repeatability, and low efficiency in traditional manual identification and counting are solved. The entire process is automated and intelligent, and the analysis of colony images can be completed in seconds, outputting objective and consistent morphological classification and counting results, which greatly improves detection efficiency and the reliability of results. A target detection model based on deep learning is used as the core recognition engine, and it is specially trained using a dataset of Bacillus subtilis colony images including multiple morphological categories. This solves the technical problem that traditional manual methods and general image processing algorithms cannot effectively learn and recognize the complex and diverse specific morphological features of Bacillus subtilis, and achieves accurate and automatic identification and classification of multiple colony morphologies. Its recognition accuracy is significantly better than that of traditional methods.

[0026] In practical implementation, the colony detection model is the YOLO series model. Due to its superior real-time detection performance and end-to-end target localization capabilities, the YOLO series model has become a highly promising technical tool in the field of intelligent microbial image recognition. Specifically, it can be the YOLOv11 model. YOLOv11 adopts a single-stage detection architecture, treating the target detection task as a unified regression problem. While ensuring high accuracy, it significantly improves detection speed, enabling real-time localization, classification, and confidence assessment of multiple targets in an image. Through multi-layer convolution and feature pyramid structures, this model can effectively extract key information such as the target's edge contour, texture distribution, and shape features. It is suitable for analyzing microbial colony images with diverse morphologies and uneven distribution. Furthermore, YOLOv11 has excellent model compression and deployment capabilities, and can be flexibly adapted to microscope image acquisition equipment, flatbed scanning systems, or automated detection platforms, demonstrating strong engineering and practical application potential.

[0027] In practice, various morphological categories are identified, including mucous-like (Y), smooth without wrinkles and protrusions (P), colonies with a ring-like structure in the center (H), and colonies with wrinkles and protrusions throughout (Z). By establishing a standardized visual definition and annotation system for these four key morphologies, the deep learning model can learn and capture subtle morphological features at different differentiation stages. This enables the automated and objective identification and statistical analysis of the multimorphic differentiation trajectories of Bacillus subtilis colonies, especially KCXM colonies, providing a quantifiable and efficient analytical tool for studying its cell differentiation mechanisms.

[0028] In specific implementation, such as Figure 3 The image shown is a comparison of the colony recognition performance between expert manual annotation and the colony detection model of this application. Figure 3 The left side shows the colonies manually labeled by experts, and the right side shows the colony recognition of the colony detection model of this application. It can be seen that the colony recognition of the colony detection model of this application is almost consistent with the manual labeling by experts, and the colony recognition of the colony detection model of this application is high.

[0029] In one exemplary embodiment, such as Figure 4 As shown, a training method for a Bacillus subtilis colony detection model based on deep learning is provided for training the colony detection model of the above embodiment. The training method includes: S21, Obtain colony images of Bacillus subtilis with various morphological types; S22, Preprocessing and data augmentation of colony images; S23, Based on preprocessed and data-enhanced colony images, construct an labeled dataset; S24. The colony detection model is trained using the labeled dataset to obtain the trained colony detection model.

[0030] By implementing steps 21 to 24 above, a dedicated labeled dataset with multiple morphological categories was acquired and constructed, solving the previous problem of lacking high-quality, multi-morphological Bacillus subtilis strain image datasets and laying a solid data foundation for high-precision recognition. Furthermore, by integrating morphological prior knowledge into the deep learning model, the technical bottleneck that general object detection models cannot be directly applied to the fine morphological recognition of specific microorganisms was solved, realizing end-to-end automated construction from raw images to dedicated, high-performance colony detection models.

[0031] In specific implementation, the colony image in step 21 above can be from a strain of Bacillus subtilis KCXM screened in the laboratory, which can spontaneously produce four types of colonies in the culture medium. The nutrient composition of the culture medium can be 0.1% yeast extract, 0.3% beef extract, 1% peptone, and 1% glucose, such as... Figure 5 The four morphological categories of colonies shown in the figure are: mucous (Y), smooth without wrinkles and protrusions (P), colonies with a ring in the middle (H), and colonies with wrinkles and protrusions throughout (Z). Bacillus subtilis KCXM was cultured under uniform shaking conditions: 180 rpm, 37°C, and 24 h. The colonies were then spread on appropriate solid agar plates. Images of the four morphological categories were captured using a camera, with each image measuring 1280 × 1280 mm. The total colony image dataset contained at least 300 high-resolution images.

[0032] In practice, preprocessing of colony images includes image denoising, illumination correction, background removal, and extraction of colony regions using image segmentation algorithms. Data augmentation of colony images includes image rotation, flipping, brightness adjustment, and image size standardization. Thus, by preprocessing and augmenting colony images, the technical challenges of poor model generalization and overfitting caused by unstable imaging conditions and limited sample size are solved. This achieves standardization and enrichment of training data, significantly improving the robustness and recognition stability of the final model in complex real-world scenarios.

[0033] In practical implementation, the colony images in the model training dataset can be labeled using the labellimg image annotation tool to annotate colonies of different morphological categories. Expert-level manual annotation is employed to ensure the accuracy of bounding boxes and category labels. The annotation of colony images in the dataset can be performed as follows: Figure 6 As shown.

[0034] In the specific implementation site, the effective data of the four morphological categories of KCXM strain at different differentiation stages were finally divided into training set (70%), validation set (15%) and test set (15%).

[0035] In practical implementation, the colony detection model is trained using transfer learning, fine-tuning the model with pre-trained model weights. Specifically, the pre-trained model can be a COCO pre-trained model, set to train for 300 iterations, with a batch size of 8 and an initial learning rate of 0.01. The AdamW optimizer and cosine annealing learning rate scheduling are used to calculate mAP@0.5 (mean precision calculated under an IoU threshold of 0.5), Precision, and Recall in real time during training. Automatic early stopping and model checkpoint saving are implemented to ultimately obtain the optimal model weights for Bacillus subtilis KCXM colony identification. Therefore, by adopting a transfer learning strategy based on the COCO dataset, combined with the AdamW optimizer, cosine annealing learning rate scheduling, and automated training monitoring, the core technical challenges of training deep learning models from scratch—such as the large amount of data required, long training cycles, and susceptibility to overfitting or convergence difficulties on small-scale specialized datasets—are effectively solved.

[0036] In an exemplary embodiment, a training method for the aforementioned deep learning-based Bacillus subtilis colony detection model is provided to train YOLO series models, and the results are compared. The models compared include YOLOv8, YOLOv9, YOLOv10, YOLOv11, and YOLOv12 from the YOLO series. Performance comparison graphs of these different models are shown below. Figure 7 As shown, it can be seen that YOLOv11 has the best performance. Figure 7 In this context, mAP@0.5 represents the mean accuracy calculated under an IoU threshold of 0.5, mAP@0.5:0.95 represents the average of the mean accuracy calculated under multiple IoU thresholds from 0.5 to 0.95, FPS represents the number of images that the model can process and recognize per second, and Model size represents the model size.

[0037] In one exemplary embodiment, such as Figure 8 As shown, a deep learning-based Bacillus subtilis colony morphology recognition and counting device is provided, including: an acquisition module 31, an input module 32, a morphology result output module 33, and a counting result output module 34.

[0038] The acquisition module 31 is configured to acquire colony images of Bacillus subtilis to be identified.

[0039] Input module 32 is configured to input the colony image of Bacillus subtilis into a trained colony detection model.

[0040] The morphological result output module 33 is configured to output the morphological recognition results of the colony image by the colony detection model, the morphological recognition results including the colony bounding box, morphological classification and confidence score.

[0041] The counting result output module 34 is configured to perform statistical analysis of the bounding boxes of all colonies in the morphological recognition results, obtain the colony count results, and output them.

[0042] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 9 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs in the non-volatile storage media to run. The database stores images of Bacillus subtilis colonies to be identified. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When executed by the processor, the computer program implements a deep learning-based method for identifying and counting Bacillus subtilis colony morphology.

[0043] Those skilled in the art will understand that Figure 9 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0044] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0045] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0046] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0047] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0048] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0049] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0050] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0051] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for identifying and counting Bacillus subtilis colony morphology based on deep learning, characterized in that, include: Obtain colony images of the Bacillus subtilis strain to be identified; The colony images of Bacillus subtilis are input into the trained colony detection model; The output is the morphological recognition result of the colony image by the colony detection model, and the morphological recognition result includes the colony bounding box, morphological classification and confidence score; The bounding boxes of all colonies in the morphological recognition results are counted to obtain the colony count results and output them. The colony detection model is based on a deep learning-based target detection model and is trained using a dataset of Bacillus subtilis colony images of various morphological categories.

2. The method for identifying and counting Bacillus subtilis colony morphology based on deep learning according to claim 1, characterized in that, The colony detection model is the YOLO series model.

3. The method for identifying and counting Bacillus subtilis colony morphology based on deep learning according to claim 2, characterized in that, The YOLO series model is the YOLOv11 model.

4. The method for identifying and counting Bacillus subtilis colony morphology based on deep learning according to claim 1, characterized in that, The various morphological categories include mucous-like, smooth without wrinkles or protrusions, colonies with a ring in the middle, and colonies with wrinkles and protrusions throughout.

5. A training method for a deep learning-based Bacillus subtilis colony detection model, used to train the colony detection model according to any one of claims 1-4, characterized in that, include: Obtain colony images of Bacillus subtilis with various morphological types; The colony images are preprocessed and data augmented; Based on the preprocessed and data-enhanced colony images, a labeled dataset is constructed; The colony detection model is trained using the labeled dataset to obtain the trained colony detection model.

6. The training method for the Bacillus subtilis colony detection model based on deep learning according to claim 5, characterized in that, The colony detection model is trained using transfer learning, and the pre-trained model weights are used to fine-tune the colony detection model.

7. The training method for the deep learning-based Bacillus subtilis colony detection model according to claim 5, characterized in that, The preprocessing methods for the colony images include image denoising, illumination correction, background removal, and extracting colony regions using image segmentation algorithms.

8. The training method for the Bacillus subtilis colony detection model based on deep learning according to claim 5, characterized in that, The methods for data augmentation of colony images include: image rotation, flipping, brightness adjustment, and image size standardization.

9. A device for recognizing and counting Bacillus subtilis colony morphology based on deep learning, characterized in that, include: The acquisition module is configured to acquire colony images of Bacillus subtilis to be identified; The input module is configured to input the Bacillus subtilis colony image into a trained colony detection model; The morphological result output module is configured to output the morphological recognition results of the colony image by the colony detection model, the morphological recognition results including the colony bounding box, morphological classification and confidence score; The counting result output module is configured to perform statistical analysis of the bounding boxes of all colonies in the morphological recognition results, obtain the colony count results, and output them.

10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the deep learning-based Bacillus subtilis colony morphology recognition and counting method as described in any one of claims 1-4.

Images