Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing fundus images using fundus image processing machine learning models. One of the methods includes obtaining a model input comprising one or more fundus images, each fundus image being an image of a fundus of an eye of a patient; processing the model input using a fundus image processing machine learning model, wherein the fundus image processing machine learning model is configured to process the model input comprising the one or more fundus image to generate a model output; and processing the model output to generate health analysis data.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing fundus images using fundus image processing machine learning models. One of the methods includes obtaining a model input comprising one or more fundus images, each fundus image being an image of a fundus of an eye of a patient; processing the model input using a fundus image processing machine learning model, wherein the fundus image processing machine learning model is configured to process the model input comprising the one or more fundus image to generate a model output; and processing the model output to generate health analysis data.

Abstract: Systems and methods for identifying visual features that influence a predictive model are provided. The technology employs an image translation function to introduce a visual feature into an image to create a modified image that can be fed to a predictive model. When the predictive model generates a different prediction for a given image than it does for a modified version of that image, the image translation function can then be used to make further modified versions that exaggerate the introduced visual feature. The technology thus aids in identifying visual features that influence the predictive model so that the model's conclusions can be understood, and so that those visual features can be further studied and tested.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing fundus images using fundus image processing machine learning models. One of the methods includes obtaining a model input comprising one or more fundus images, each fundus image being an image of a fundus of an eye of a patient; processing the model input using a fundus image processing machine learning model, wherein the fundus image processing machine learning model is configured to process the model input comprising the one or more fundus image to generate a model output; and processing the model output to generate health analysis data.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing fundus images using fundus image processing machine learning models. One of the methods includes obtaining a model input comprising one or more fundus images, each fundus image being an image of a fundus of an eye of a patient; processing the model input using a fundus image processing machine learning model, wherein the fundus image processing machine learning model is configured to process the model input comprising the one or more fundus image to generate a model output; and processing the model output to generate health analysis data.

Abstract: Camera pose optimization, which includes determining the position and orientation of a camera in three-dimensional space at different times, is improved by detecting a higher-confidence reference object in the photographs captured by the camera and using the object to increase consistency and accuracy of pose data. Higher-confidence reference objects include objects that are stationary, fixed, easily recognized, and relatively large. In one embodiment, street level photographs of a geographic area are collected by a vehicle with a camera. The captured images are geo-coded using GPS data, which may be inaccurate. The vehicle drives in a loop and captures the same reference object multiple times from the substantially same position. The trajectory of the vehicle is then closed by aligning the points of multiple images where the trajectory crosses itself. This creates an additional constraint on the pose data, which in turn improves the data's consistency and accuracy.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing fundus images using fundus image processing machine learning models. One of the methods includes obtaining a model input comprising one or more fundus images, each fundus image being an image of a fundus of an eye of a patient; processing the model input using a fundus image processing machine learning model, wherein the fundus image processing machine learning model is configured to process the model input comprising the one or more fundus image to generate a model output; and processing the model output to generate health analysis data.

Abstract: Camera pose optimization, which includes determining the position and orientation of a camera in three-dimensional space at different times, is improved by detecting a higher-confidence reference object in the photographs captured by the camera and using the object to increase consistency and accuracy of pose data. Higher-confidence reference objects include objects that are stationary, fixed, easily recognized, and relatively large. In one embodiment, street level photographs of a geographic area are collected by a vehicle with a camera. The captured images are geo-coded using GPS data, which may be inaccurate. The vehicle drives in a loop and captures the same reference object multiple times from the substantially same position. The trajectory of the vehicle is then closed by aligning the points of multiple images where the trajectory crosses itself. This creates an additional constraint on the pose data, which in turn improves the data's consistency and accuracy.

Abstract: Camera pose optimization, which includes determining the position and orientation of a camera in three-dimensional space at different times, is improved by detecting a higher-confidence reference object in the photographs captured by the camera and using the object to increase consistency and accuracy of pose data. Higher-confidence reference objects include objects that are stationary, fixed, easily recognized, and relatively large. In one embodiment, street level photographs of a geographic area are collected by a vehicle with a camera. The captured images are geo-coded using GPS data, which may be inaccurate. The vehicle drives in a loop and captures the same reference object multiple times from the substantially same position. The trajectory of the vehicle is then closed by aligning the points of multiple images where the trajectory crosses itself. This creates an additional constraint on the pose data, which in turn improves the data's consistency and accuracy.

Abstract: Camera pose optimization, which includes determining the position and orientation of a camera in three-dimensional space at different times, is improved by detecting a higher-confidence reference object in the photographs captured by the camera and using the object to increase consistency and accuracy of pose data. Higher-confidence reference objects include objects that are stationary, fixed, easily recognized, and relatively large. In one embodiment, street level photographs of a geographic area are collected by a vehicle with a camera. The captured images are geo-coded using GPS data, which may be inaccurate. The vehicle drives in a loop and captures the same reference object multiple times from the substantially same position. The trajectory of the vehicle is then closed by aligning the points of multiple images where the trajectory crosses itself. This creates an additional constraint on the pose data, which in turn improves the data's consistency and accuracy.