Abstract: An automated tuning service is used to automatically tune, or modify, the operational parameters of a large-scale cloud infrastructure. The tuning service performs automated and fully data/model-driven configuration based from learning various real-time performance of the cloud infrastructure. Such performance is identified through monitoring various telemetric data of the cloud infrastructure. The tuning service leverages a mix of domain knowledge and principled data-science to capture the essence of our cluster dynamic behavior in a collection of descriptive machine learning (ML) models. The ML models power automated optimization procedures for parameter tuning, and inform administrators in most tactical and strategical engineering/capacity decisions (such as hardware and datacenter design, software investments, etc.).

Abstract: An image processing apparatus (100) includes: an image estimation unit (124) that estimates a removal image obtained by removing, from a visible image, a scatter component of sunlight scattered when passing through an atmospheric layer, on the basis of the visible image and an infrared image captured simultaneously with the visible image.

Abstract: An automated tuning service is used to automatically tune, or modify, the operational parameters of a large-scale cloud infrastructure. The tuning service performs automated and fully data/model-driven configuration based from learning various real-time performance of the cloud infrastructure. Such performance is identified through monitoring various telemetric data of the cloud infrastructure. The tuning service leverages a mix of domain knowledge and principled data-science to capture the essence of our cluster dynamic behavior in a collection of descriptive machine learning (ML) models. The ML models power automated optimization procedures for parameter tuning, and inform administrators in most tactical and strategical engineering/capacity decisions (such as hardware and datacenter design, software investments, etc.).

Abstract: Methods, systems and computer program products are provided for predicting runtime variation in big data analytics. Runtime probability distributions may be predicted for proposed computing jobs. A predictor may classify proposed computing jobs based on multiple runtime probability distributions that represent multiple clusters of runtime probability distributions for multiple executed recurring computing job groups. Proposed computing jobs may be classified as delta-normalized runtime probability distributions and/or a ratio-normalized runtime probability distributions. Sources of runtime variation may be identified with a quantitative contribution to predicted runtime variation. A runtime probability distribution editor may indicate modifications to sources of runtime variation in a proposed computing job and/or predict reductions in predicted runtime variation provided by modifications to a proposed computing job.

Abstract: A pixel addition processing unit 35 adds pupil division pixel signals with a plurality of different addition patterns in pupil units to generate a pixel addition image for each addition pattern. Further, the pixel addition processing unit 35 may use a pupil division pixel signal in which a sensitivity difference between pupil division pixels has been corrected, and the addition pattern may be set on the basis of image characteristic information calculated by using the pupil division pixel signal. A super-resolution processing unit 36 performs super-resolution processing using a pixel addition image signal generated for each addition pattern by the pixel addition processing unit 35 to generate an image signal with a higher resolution than that of the pixel addition image signal. A high-resolution image can be generated using pupil division.

Abstract: An imaging element of an imaging unit 24 divides the exit pupil of an imaging optical system 21 into a plurality of regions and generates a pixel signal for each region. An optical axis position adjustment unit 23 adjusts the optical axis position of the imaging optical system with respect to the imaging element. A control unit 26 calculates a parallax on the basis of the pixel signal for each region after the pupil division and performs focus control of the imaging optical system 21. The control unit 26 also moves the optical axis position using the optical axis position adjustment unit 23, and generates, using the imaging element, pixel signals indicating the same subject region in the plurality of regions after the pupil division. An image processing unit 25 performs binning of a plurality of pixel signals indicating the same subject region generated by moving the optical axis position to generate a high-resolution captured image.

Abstract: Methods, systems, apparatuses, and computer-readable storage mediums described herein are directed to determining and recommending an optimal compute resource configuration for a cloud-based resource (e.g., a server, a virtual machine, etc.) for migrating a customer to the cloud. The embodiments described herein utilize a statistically robust approach that makes recommendations that are more flexible (elastic) and account for the full distribution of the amount of resource usage. Such an approach is utilized to develop a personalized rank of relevant recommendations to a customer. To determine which compute resource configuration to recommend to the customer, the customer’s usage profile is matched to a set of customers that have already migrated to the cloud. The compute resource configuration that reaches the performance most similar to the performance of the configurations utilized by customers in the matched set is recommended to the user.

Abstract: There is provided an imaging device (10) including: an imaging module (100) including an image sensor (130) in which a plurality of pixels for converting light into an electric signal is arranged; a drive unit (140) that moves a part of the imaging module in a manner that the image sensor can sequentially acquire a reference image under a predetermined pixel phase, a plurality of generation images, and a detection image under the predetermined pixel phase in this order; and a detection unit (220) that detects a moving subject based on a difference between the reference image and the detection image.

Abstract: Techniques are described herein that are capable of selecting checkpoints of a database job. For instance, at compile time, temporal indicators associated with the query plans of the database job are determined. Each temporal indicator indicates first and second subsets of stages of the respective query plan. Values of attributes of each stage in at least each first subset are predicted using a machine learning technique. At the compile time, candidate stage(s) for each query plan are identified based on the respective candidate stage being a child of stage(s) in the corresponding second subset or not being a child of another stage in the respective query plan. The candidate stage(s) for each query plan are selectively chosen as respective checkpoint(s) based on whether the values of the attributes of each stage in at least the first subset of the stages of the respective query plan satisfy one or more criteria.

Abstract: An imaging element of an imaging unit 24 divides the exit pupil of an imaging optical system 21 into a plurality of regions and generates a pixel signal for each region. An optical axis position adjustment unit 23 adjusts the optical axis position of the imaging optical system with respect to the imaging element. A control unit 26 calculates a parallax on the basis of the pixel signal for each region after the pupil division and performs focus control of the imaging optical system 21. The control unit 26 also moves the optical axis position using the optical axis position adjustment unit 23, and generates, using the imaging element, pixel signals indicating the same subject region in the plurality of regions after the pupil division. An image processing unit 25 performs binning of a plurality of pixel signals indicating the same subject region generated by moving the optical axis position to generate a high-resolution captured image.

Abstract: An automated tuning service is used to automatically tune, or modify, the operational parameters of a large-scale cloud infrastructure. The tuning service performs automated and fully data/model-driven configuration based from learning various real-time performance of the cloud infrastructure. Such performance is identified through monitoring various telemetric data of the cloud infrastructure. The tuning service leverages a mix of domain knowledge and principled data-science to capture the essence of our cluster dynamic behavior in a collection of descriptive machine learning (ML) models. The ML models power automated optimization procedures for parameter tuning, and inform administrators in most tactical and strategical engineering/capacity decisions (such as hardware and datacenter design, software investments, etc.).

Abstract: A pixel addition processing unit 35 adds pupil division pixel signals with a plurality of different addition patterns in pupil units to generate a pixel addition image for each addition pattern. Further, the pixel addition processing unit 35 may use a pupil division pixel signal in which a sensitivity difference between pupil division pixels has been corrected, and the addition pattern may be set on the basis of image characteristic information calculated by using the pupil division pixel signal. A super-resolution processing unit 36 performs super-resolution processing using a pixel addition image signal generated for each addition pattern by the pixel addition processing unit 35 to generate an image signal with a higher resolution than that of the pixel addition image signal. A high-resolution image can be generated using pupil division.

Abstract: Provided is an image processing device that includes a wide image capturing unit which captures an image of a wide viewing angle, a tele image capturing unit which captures an image of a narrow viewing angle, and an image processing unit which receives an input of the captured images of the respective imaging units and executes signal processing. In a case where a wide conversion lens is attached to the tele image capturing unit, the image processing unit executes any one of a high-sensitivity image generation processing, a HDR image generation processing, or a high-resolution image generation processing based on the captured images of the wide image capturing unit and the tele image capturing unit according to a situation at the time of imaging with a wide conversion lens attachment detection unit which detects whether the wide conversion lens is attached.

Abstract: Techniques are described herein that are capable of selecting checkpoints of a database job. For instance, at compile time, temporal indicators associated with the query plans of the database job are determined. Each temporal indicator indicates first and second subsets of stages of the respective query plan. Values of attributes of each stage in at least each first subset are predicted using a machine learning technique. At the compile time, candidate stage(s) for each query plan are identified based on the respective candidate stage being a child of stage(s) in the corresponding second subset or not being a child of another stage in the respective query plan. The candidate stage(s) for each query plan are selectively chosen as respective checkpoint(s) based on whether the values of the attributes of each stage in at least the first subset of the stages of the respective query plan satisfy one or more criteria.

Abstract: Provided are a signal processing apparatus, a signal processing method, and an imaging apparatus for estimating depth with a high degree of accuracy from video signals of a plurality of cameras. The signal processing apparatus includes a detection unit that detects depth from video captured by a plurality of cameras, an estimation unit that estimates camera motion by detecting a camera position, a prediction unit that predicts present depth from camera motion and depth obtained previously, and a synthesis unit that synthesizes depth detected by the detection unit and depth predicted by the prediction unit on the basis of a result of estimating the amount of blur that occurs in each image captured by cameras. The amount of blur can be estimated using camera motion, previous depth information, and shutter time.

Abstract: To realize a device and a method capable of generating a high-quality image similar to that of a same viewing angle compound eye camera by using captured images of a tele/wide compound eye camera. A wide image capturing unit which captures an image of a wide viewing angle, a tele image capturing unit which captures an image of a narrow viewing angle, and an image processing unit which receives an input of the captured images of the respective imaging units and executes signal processing are included.

Abstract: Provided are a signal processing apparatus, a signal processing method, and an imaging apparatus for estimating depth with a high degree of accuracy from video signals of a plurality of cameras. The signal processing apparatus includes a detection unit that detects depth from video captured by a plurality of cameras, an estimation unit that estimates camera motion by detecting a camera position, a prediction unit that predicts present depth from camera motion and depth obtained previously, and a synthesis unit that synthesizes depth detected by the detection unit and depth predicted by the prediction unit on the basis of a result of estimating the amount of blur that occurs in each image captured by cameras. The amount of blur can be estimated using camera motion, previous depth information, and shutter time.

Abstract: There is provided an apparatus and a method that perform a process of improving the quality of a low-quality image such as a far-infrared image. The apparatus includes an image correction unit that repeatedly performs an image correction process using a plurality of processing units in at least two stages. The image correction unit inputs a low-quality image to be corrected and a high-quality image which is a reference image. Each of the processing units in each stage performs a correction process for the low-quality image, using a class correspondence correction coefficient corresponding to a feature amount extracted from a degraded image of the high-quality image. A processing unit in a previous stage performs the correction process, using a class correspondence correction coefficient corresponding to a feature amount extracted from an image which has a higher degradation level than that in a processing unit in a subsequent stage.

Abstract: Effective image correction processing for reducing flicker is executed according to characteristics of images, and an image to be displayed in a liquid crystal display apparatus is generated. Characteristic amount change rate data that is a change rate between a characteristic amount of a sample image and a characteristic amount of a sample image output to a liquid crystal display device is acquired in advance and stored in a storage unit. The correction parameter for reducing flicker is calculated on the basis of the characteristic amount of the image to be corrected and the characteristic amount change rate data of the sample images stored in the storage unit. The correction processing to which the calculated correction parameter has been applied is executed for the image to be corrected to generate a display image. As the characteristic amount, for example, the interframe luminance change amount, the interline luminance conversion amount, or the interframe motion vector is used.

Abstract: A coefficient calculator calculates a correlation coefficient representing a correlation between a normal frame, which is imaged in a state in which normal light is applied to a subject, and a special frame, which is imaged in a state in which special light is applied to the subject; and a processing unit that applies image processing to the special frame so that a part in which the correlation coefficient is high and a part in which the correlation coefficient is low are displayed differently in the special frame. Further provided are: a first determination unit that calculates, for each pixel or each area of the normal frame, a first determination value representing a probability that a predetermined site is imaged; and a second determination unit that calculates, for each pixel or each area of the special frame, a second determination value representing a probability that a predetermined site is imaged.