Automated spermatozoa candidate identification
A method comprising: receiving image data associated with a plurality of semen samples; at a training stage, training a machine learning model on a training set comprising: (i) said image data, and (ii) labels associated with a qualitative assessment of each of one or more individual spermatozoa in said semen samples; and applying said trained machine learning model to target image data associated with a target semen sample, to identify one or more spermatozoa in said target sample as candidates for an ART procedure.
1. A system comprising:
at least one hardware processor; and
a non-transitory computer-readable storage medium having stored thereon program instructions, the program instructions executable by the at least one hardware processor to:
receive series of images associated with a target semen sample,
perform an image processing stage to detect one or more regions-of-interest (ROI) in the received series of images, said ROI associated with an individual spermatozoon in the target semen sample;
apply a trained machine learning model to the detected one or more ROI, to calculate an estimated success rate of fertility of said individual spermatozoon in an insemination procedure; said trained machine learning model being trained on a training set comprising:
(i) series of images associated with a plurality of semen samples, and
(ii) labels associated with qualitative parameters associated with an estimated success rate of fertility of each of one or more individual spermatozoa in said plurality of semen samples in the insemination procedure, said qualitative parameters comprising morphology parameters and motility parameters; and
identify said individual spermatozoon as a candidate for the insemination procedure, based on the estimated success rate.
2. The system of claim 1 , wherein said motility parameters are selected from a group consisting of: detected motility, progressive motility, and linear motility, and wherein said morphology parameters are selected from a group consisting of: base morphology, head morphology, and a presence and location of one or more vacuoles.
3. The system of claim 1 , wherein said comprises at least one of: image data cleaning, image data normalization, identification of said individual spermatozoa in said series of images associated with said target semen sample.
4. The system of claim 3 , further comprising tracking of individual spermatozoon in said series of images associated with said target semen sample, wherein said tracking comprises identifying coordinates for said individual spermatozoon in said series of images associated with the target semen sample.
5. The system of claim 4 , wherein said tracking further comprises operating a retrieval device to retrieve said individual spermatozoon from said target semen sample, and wherein said retrieval is based on said identified coordinates.
6. The system of claim 1 , wherein said instructions are further executable to perform a feature selection stage, and wherein said feature selection comprises selection of one or more subsets of data points within the image data.
7. The system of claim 1 , wherein said series of images comprise, with respect to each of said semen samples, at least one of: a video segment, a streamed video segment, and a real-time video segment.
8. The system of claim 1 , wherein said identification comprises a confidence score assigned to said individual spermatozoon in said target semen sample.
9. A method comprising:
receiving series of images associated with a target sample semen;
performing an image processing stage to detect one or more regions-of-interest (ROI) in the received series of images, said ROI associated with an individual spermatozoon in the target semen sample;
applying a trained machine learning model to the detected one or more ROI, to calculate an estimated success rate of fertility of said individual spermatozoon in an insemination procedure, said trained machine learning model being trained on a training set comprising:
(i) series of images associated with a plurality of semen samples, and
(ii) labels associated with qualitative parameters associated with an estimated success rate of fertility of each of one or more individual spermatozoa in said plurality of semen samples in the insemination procedure, said qualitative parameters comprising morphology parameters and motility parameters; and
identifying said individual spermatozoon as a candidate for an insemination procedure, based on the estimated success rate.
10. The method of claim 9 , wherein said motility parameters are selected from a group consisting of: detected motility, progressive motility, and linear motility, and wherein said morphology parameters are selected from a group consisting of: base morphology, head morphology, and a presence and location of one or more vacuoles.
11. The method of claim 9 , wherein said image processing stage comprises at least one of: image data cleaning, image data normalization, identification of said individual spermatozoa in said series of images associated with said target semen sample.
12. The method of claim 11 , further comprising tracking of individual spermatozoon in said series of images associated with said target semen sample, wherein said tracking comprises identifying coordinates for said individual spermatozoon from said target semen sample.
13. The method of claim 12 , wherein said tracking further comprises operating a retrieval device to retrieve said individual spermatozoon from said target sample, and wherein said retrieval is based on said identified coordinates.
14. The method of claim 9 , further comprising performing a feature selection stage, wherein said feature selection comprises selection of one or more subsets of data points within the image data.
15. The method of claim 9 , wherein said series of images comprise, with respect to each of said semen samples, at least one of: a video segment, a streamed video segment, and a real-time video segment.
16. The method of claim 9 , wherein said identifying comprises assigning a confidence score to said individual spermatozoon in said target semen sample.
17. A computer program product comprising a non-transitory computer-readable storage medium having program instructions embodied therewith, the program instructions executable by at least one hardware processor to:
receive series of images associated with a target semen sample;
perform an image processing stage to detect one or more regions-of-interest (ROI) in the received series of images, said ROI associated with an individual spermatozoon in the target semen sample;
apply a trained machine learning model to the detected one or more ROI, to calculate an estimated success rate of fertility of said individual spermatozoon in an insemination procedure, said trained machine learning model being trained on a training set comprising:
(i) series of images associated with a plurality of semen samples, and
(ii) labels associated with qualitative parameters associated with an estimated success rate of fertility of each of one or more individual spermatozoa in said plurality of semen samples in the insemination procedure, said qualitative parameters comprising morphology parameters and motility parameters; and
identify said individual spermatozoon as a candidates for an insemination procedure, based on the estimated success rate.
18. The computer program product of claim 17 , wherein said motility parameters are selected from a group consisting of: detected motility, progressive motility, and linear motility, and wherein said morphology parameters are selected from a group consisting of: base morphology, head morphology, and a presence and location of one or more vacuoles, and
wherein said image processing stage comprises at least one of: image data cleaning, image data normalization, identification of individual spermatozoa in said series of images associated with said target semen sample, and tracking of individual spermatozoon in said series of images associated with said target semen sample.