Abstract: A polarization imaging unit 20 acquires a polarized image including polarization pixels with a plurality of polarization directions. An information compression unit 40 sets reference image information based on polarized image information of reference polarization pixels with at least a plurality of polarization directions in the polarized image and generates difference information between the polarized image information of each of polarization pixels different from the reference polarization pixels in the polarized image and the reference image information. In addition, the information compression unit 40 reduces the amount of information of the difference information generated for each of the polarization pixels with the plurality of polarization directions to generate compressed image information including the reference image information and the difference information with the reduced amount of information.

Abstract: An information processing apparatus according to an embodiment of the present technology includes a detection unit, an estimation unit, and a prediction unit. The detection unit detects a target object from an input image. The estimation unit estimates a posture of the detected target object. The prediction unit predicts an action of the target object on a basis of the estimated posture.

Abstract: Methods, non-transitory machine readable media, and computing devices that dynamically throttle non-priority workloads to satisfy minimum throughput service level objectives (SLOs) are disclosed. With this technology, a determination is made when a number of detection intervals with a violation within a detection window exceeds a threshold, when a current one of the detection intervals is outside an observation area. The detection intervals are identified a violated based on an average throughput for priority workloads within the detection intervals exceeding a minimum throughput SLO. A throttle is then set to rate-limit non-priority workloads, when the number of violated detection intervals within the detection window exceeds the threshold.

Abstract: A polarization imaging unit generates, as polarization information, a polarization image including pixels in a plurality of polarization directions. An interpolation processing unit of an information processor performs interpolation processing using the polarization image obtained from the polarization imaging unit, and generates a polarization image for each polarization method. A polarization degree calculation unit calculates, for example, a polarization degree for each pixel as object surface information on the basis of the polarization image for each polarization method. A noise amount calculation unit calculates a noise amount for each pixel on the basis of the polarization image for each polarization method and the like.

Abstract: Methods, non-transitory machine readable media, and computing devices that dynamically throttle non-priority workloads to satisfy minimum throughput service level objectives (SLOs) are disclosed. With this technology, a determination is made when a number of detection intervals with a violation within a detection window exceeds a threshold, when a current one of the detection intervals is outside an observation area. The detection intervals are identified a violated based on an average throughput for priority workloads within the detection intervals exceeding a minimum throughput SLO. A throttle is then set to rate-limit non-priority workloads, when the number of violated detection intervals within the detection window exceeds the threshold.

Abstract: A damage reduction device according to an embodiment of the present technology includes an input unit, a prediction unit, a recognition unit, and a determination unit. The input unit inputs status data regarding a status in a moving direction of a moving body apparatus. The prediction unit predicts a collision with an object in the moving direction on the basis of the status data. The recognition unit recognizes whether the object includes a person. The determination unit determines, when the collision with the object is predicted and it is recognized that the object includes a person, a steering direction of the moving body apparatus in which a collision with the person is avoidable, on the basis of the status data.

Abstract: The invalid pixel detection unit 32 detects an invalid pixel from a non-polarized image and a plurality of polarized images for different polarization directions obtained by performing image pickup using the polarization image pickup unit 20. For example, the invalid pixel detection unit 32 detects a saturated pixel having a pixel value larger than a preset saturation detection threshold and a black-crushed pixel having a pixel value smaller than a preset black crushing detection threshold as invalid pixels from a non-polarized image. The polarization information generation unit 33 performs processing of generating polarization information on the basis of the non-polarized image and polarized images and switches the processing of generating the polarization information depending on the detection result of the invalid pixels in the invalid pixel detection unit 32, so as to generate the polarization information without using the invalid pixels. Correct polarization information can be acquired.

Abstract: The present technology relates to a learning device, a generation method, an inference device, an inference method, and a program that enable learning data suitable for learning to be selected without a manual operation and enable learning of an inference model to be efficiently performed by using the selected learning data. In a learning device according to one aspect of the present technology, on the basis of a learning data group including learning data having a correct answer and a processing target data group including processing target data for learning, the processing target data for learning having no correct answer and corresponding to data to be processed at the time of inference, the learning device selects, from the learning data group, the learning data suitable for learning of an inference model used at the time of inference, and outputs the selected learning data together with the inference model obtained by performing learning using the selected learning data.

Abstract: A same region detector detects a same region of a processing target for each of a plurality of different viewpoints from polarization images in a plurality of polarization directions acquired for each of the viewpoints. The polarization images in the plurality of polarization directions acquired for each of the plurality of different viewpoints are, for example, polarization images acquired by imaging over a period of a plurality of frames in which a positional relationship between the processing target and a polarization image acquisition unit that acquires the polarization images changes. A polarization degree calculation unit calculates a polarization degree of the same region for each of the viewpoints on the basis of the polarization images in the plurality of polarization directions.

Abstract: A polarization rendering setting unit 30 sets a plurality of parameters to be used for generating a polarized rendered image of an abnormality detection target. A polarized rendered image generation unit 40 generates the polarized rendered image of the abnormality detection target on the basis of the parameters set by the polarization rendering setting unit 30. An abnormality detection unit 50 detects an abnormal region of the abnormality detection target on the basis of a difference between a polarized captured image acquired by imaging the abnormality detection target and the polarized rendered image generated by the polarized rendered image generation unit 40. Abnormalities that are difficult to detect on the basis of luminance information and color information becomes able to be detected on the basis of polarized images.

Abstract: A normal candidate information generation unit of an information processor generates normal candidate information for each pixel indicating, for example, a zenith angle, or an azimuth angle, or a zenith angle and an azimuth angle, on the basis of a polarization image in a plurality of polarization directions obtained by a polarization imaging unit. The in-plane pixel selection unit selects a plurality of pixels indicating the plane to be observed in the polarization image. A normal calculation unit calculates a normal of the plane to be observed on the basis of the normal candidate information of the pixels selected by the in-plane pixel selection unit.

Abstract: A polarized image acquisition section 11a acquires a polarized image of a target object having one or more polarization directions. A polarization parameter acquisition section 12-1 calculates the average brightness ? of a polarization model on the basis of a non-polarized image subjected to sensitivity correction. Further, the polarization parameter acquisition section 12-1 calculates the amplitude ? of the polarization model on the basis of the calculated average brightness ?, pre-stored information regarding the zenith angle ? of the normal line of the target object, a refractive index r, and reflectance property information indicative of whether a subject is diffuse reflection or specular reflection.

Abstract: Provided is a polarization imaging unit that acquires a polarization image using a lens. In a parameter storage unit of an information processing unit, parameters related to a change in polarization state due to the lens which are estimated in advance from polarization state information acquired by imaging a light source in a predetermined polarization state using the lens and polarization state information indicating the polarization state of the light source, for example, change parameters indicating the change in polarization state due to the lens or correction parameters for correcting the change in polarization state due to the lens are stored. A polarization state correction unit of the information processing unit corrects the change in polarization state due to the lens, using the parameters stored in advance in the parameter storage unit.

Abstract: An unpolarized component image generating section generates unpolarized component image signals of specific and non-specific colors. A polarized component image generating section generates polarized component image signals of the specific and non-specific colors. A specific color polarization property detecting section detects polarization properties of the specific color using pixel signals of polarized or unpolarized pixels of the specific color having at least three polarization directions and pixel signals of polarized pixels having two polarization directions.

Abstract: The invalid pixel detection unit 32 detects an invalid pixel from a non-polarized image and a plurality of polarized images for different polarization directions obtained by performing image pickup using the polarization image pickup unit 20. For example, the invalid pixel detection unit 32 detects a saturated pixel having a pixel value larger than a preset saturation detection threshold and a black-crushed pixel having a pixel value smaller than a preset black crushing detection threshold as invalid pixels from a non-polarized image. The polarization information generation unit 33 performs processing of generating polarization information on the basis of the non-polarized image and polarized images and switches the processing of generating the polarization information depending on the detection result of the invalid pixels in the invalid pixel detection unit 32, so as to generate the polarization information without using the invalid pixels. Correct polarization information can be acquired.

Abstract: A quantization range setting unit (31) of an encoder (30) sets a quantization range for each Stokes parameter of a Stokes vector acquired from a Stokes vector calculation unit (20). The quantization range is set for the Stokes parameter indicating intensity, and is then set for the other Stokes parameters. A quantization unit (32) calculates the Stokes parameter indicating intensity as a predetermined quantization bit number, and calculates the quantization bit numbers of the other Stokes parameters on the basis of the predetermined quantization bit number and the quantization ranges set for the respective Stokes parameters. The quantization unit (32) performs a quantization process on the Stokes parameters on the basis of the quantization ranges and the quantization bit numbers, to generate quantized polarization information. A decoder (40) performs inverse quantization compatible with the encoder 30 on the quantized polarization information, and generates the Stokes vectors before quantization.

Abstract: A third imaging unit including a pixel not having a polarization characteristic is interposed between a first imaging unit and a second imaging unit including a pixel having a polarization characteristic for each of a plurality of polarization directions. A depth map is generated from a viewpoint of the first imaging unit by matching processing using a first image generated by the first imaging unit and a second image generated by the second imaging unit. A normal map is generated on the basis of a polarization state of the first image. Integration processing of the depth map and the normal map is performed and a depth map with a high accuracy is generated. The depth map generated by the map integrating unit is converted into a map from a viewpoint of the third imaging unit, and an image free from deterioration can be generated.

Abstract: Methods, non-transitory machine readable media, and computing devices that dynamically throttle non-priority workloads to satisfy minimum throughput service level objectives (SLOs) are disclosed. With this technology, a determination is made when a number of detection intervals with a violation within a detection window exceeds a threshold, when a current one of the detection intervals is outside an observation area. The detection intervals are identified a violated based on an average throughput for priority workloads within the detection intervals exceeding a minimum throughput SLO. A throttle is then set to rate-limit non-priority workloads, when the number of violated detection intervals within the detection window exceeds the threshold.

Abstract: An imaging apparatus (10) includes an imaging unit (11) that has a first pixel block and a second pixel block provided in a pixel region, the first pixel block including polarization pixels with two different polarization directions and a non-polarization pixel, the second pixel block including polarization pixels with a plurality of polarization directions including a different polarization direction from the first pixel block. An image processing apparatus (20) acquires a captured image generated by the imaging apparatus (10), and a correction coefficient calculation unit (40) calculates a correction coefficient that absorbs a sensitivity difference between polarization pixels and non-polarization pixels, on the basis of an image of the second pixel block.

Abstract: A polarization imaging section 20 includes polarized pixels in each of a plurality of polarization directions. The polarization imaging section 20 includes a polarizer. The polarization imaging section 20 outputs image signals of a polarized image to a defect detecting section 35 of an image processing section 30. In a case where a difference between a pixel value of a target polarized pixel generated by the polarization imaging section and a pixel value of the target polarized pixel estimated from polarization characteristics corresponding to pixel values of peripheral pixels in a polarization direction different from a polarization direction of the target polarized pixel is greater than a predetermined allowable range, the defect detecting section 35 determines that the target polarized pixel is a defective pixel. Therefore, it is possible to detect a defect of a pixel in the polarization imaging section that generates the polarized image.