Microorganic detection system using a deep learning model
Aspects of the present disclosure relate to a method of colony enumeration. The method includes identifying colony forming units of microorganisms in a combined image using a pretrained deep learning model on a colony enumeration device. The method can include providing a plurality of identification characteristics of the colony forming units to an interaction component such that the interaction component can project at least some of the plurality of identification characteristics onto the combined image.
1 . A method of colony enumeration, comprising:
identifying colony forming units of microorganisms in a combined image using a pretrained deep learning model on a colony enumeration device;
providing a plurality of identification characteristics of the colony forming units to an interaction component such that the interaction component can project at least some of the plurality of identification characteristics onto the combined image;
image capturing, with image capture circuitry, a culture device at a plurality of illumination settings to form a plurality of images of the culture device at a single incubation instance;
image enhancing, with image enhancement circuitry, the plurality of images to form a plurality of enhanced images; and
transmitting of the plurality of enhanced images to the interaction component, wherein at least two of the image capturing process, image enhancing process, or transmitting to the interaction component process occur concurrently for different image instances, wherein the combined image is formed from at least some of the plurality of enhanced images;
wherein the interaction component is configured to provide a plurality of colony forming units projected onto the combined image of the culture device within 6 seconds of starting a culture device through a feed mechanism.
2 . The method of claim 1 , wherein image enhancing some of the plurality of images occurs concurrent with image capturing the culture device at some of the plurality of illumination settings.
3 . The method of claim 1 , wherein the image enhancing process comprises flat field normalization, convolutions, sharpening, histogramic equalization, contrast enhancements, and combinations thereof.
4 . The method of claim 1 , wherein each of the plurality of illumination settings is associated with an image instance, wherein a first image instance is processed concurrently with a second image instance such that the capturing an image associated with the second image instance is triggered based on starting image enhancing on an image associated with the first image instance.
5 . The method of claim 1 , wherein the capturing the image associated with the second image instance occurs concurrently with transmitting an enhanced image associated with the first image instance.
6 . The method of claim 5 , wherein identifying colony forming units further comprises providing the enhanced image to the pretrained deep learning model, wherein at least some processes of the pretrained deep learning model are performed on the first enhanced image until a final enhanced image in a sequence of enhanced images is transmitted to the interaction component.
7 . The method of claim 1 wherein identifying colony forming units of microorganisms in the culture device comprises transmitting the combined image to the pretrained deep learning model trained to identify a colony morphology characteristic in the image, and receiving, from the pretrained deep learning model, a probability of the colony forming unit of a microorganism being present in the image based on the colony morphology characteristic.
8 . The method of claim 1 , wherein the plurality of identification characteristics are selected from: probability of the colony forming unit being a colony type, a microorganism associated with the colony forming unit, coordinates of the center point of a colony forming unit, an extent of the colony forming unit, and combinations thereof.
9 . The method of claim 1 , further comprising:
training a deep learning model with a corpus of identified colony forming units of microorganisms and colony morphology characteristics thereof to form the pretrained deep learning model.
10 . A non-transitory computer-readable storage medium including instructions that, when processed by a computer, configure the computer to perform the method of claim 1 .
11 . The method of claim 1 , wherein the plurality of images and the plurality of enhanced images comprise three images and three enhanced images, respectively, corresponding to red, blue, and green color channels and the combined images of the three enhanced images is an RGB image;
wherein the enhanced images corresponding to the red, blue, and green color channels are combined via interleaving bytewise to form a 24-bit RGB image.
12 . A colony enumeration device, comprising:
a single-board computer comprising:
neural network circuitry,
input output circuitry configured to communicatively couple to an interaction component;
a processor;
image capture circuitry communicatively coupled to the single-board computer and configured to image capture a culture device at a plurality of illumination settings to form a plurality of images at a single incubation instance, wherein the image capture circuitry comprises illumination circuitry that controls the plurality of illumination settings; and
a memory storing instructions that, when executed by the processor, configure the single-board computer to:
identify colony forming units of microorganisms in a combined image, using a pretrained deep learning model on the neural network circuitry; and
provide a plurality of identification characteristics of the colony forming units to an interaction component such that the interaction component can project at least some of the plurality of identification characteristics onto the combined image.
13 . The colony enumeration device of claim 12 ,
wherein the single-board computer comprises image enhancement circuitry configured to perform image enhancing on the plurality of images to form a plurality of enhanced images, and
wherein the memory stores instructions that, when executed by the processor, configured the input output circuitry to transmit the plurality of enhanced images to an interaction component using direct memory access, wherein at least two of the image capturing process, image enhancing process, or transmitting to the interaction component process occur concurrently for different image instances.
14 . The colony enumeration device of claim 13 , wherein the single-board computer uses overlapped interleaved processing for at least two of the image capturing process, image enhancing process, or transmitting to the interaction component process.
15 . The colony enumeration device of claim 13 , wherein the image capturing process of a subsequent image instance is triggered by the image enhancing process of a prior image instance.
16 . The colony enumeration device of claim 13 , wherein the memory stores instructions that, when executed by the processor, configured the single-board computer to form the combined image from at least some of the plurality of enhanced images that were previously transmitted.
17 . The colony enumeration device of claim 13 , further comprising:
a feed mechanism communicatively coupled to the single-board computer and configured to receive a culture device and position the culture device for image capture;
wherein the memory stores instructions that, when executed by the processor, configure the single-board computer to:
receive an indication from a feed mechanism that the culture device has been inserted and is in position; and
image capture the culture device in response to receiving the indication.
18 . A microorganic detection system comprising:
the colony enumeration device of claim 13 ; and an interaction component comprising a display device.
19 . The system of claim 18 , wherein the interaction component is configured to project at least some of the plurality of identification characteristics onto the combined image within 6 seconds of the culture device being received by a feed mechanism of the colony enumeration device.