Abstract: A check valve that can prevent sticking of a valve body and reliably perform an opening and closing action in a check vale having a simple structure and that does not use biasing means on the valve body. A check valve has support pieces of a valve body protruding in the valve body direction on a circular base seat forming an outflow window in the center inserted in a circular groove where a support member (retainer) is formed between an outflow opening and a valve seat and on an open edge of the outflow window formed in the center of the base seat, and the support pieces of the support member (retainer) that support the valve body have linear or planar support parts on the top portion for leveling at least in a radial direction of the base seat.

Abstract: A computerized method of artifact regulation in deep model training for image transformation first performs one cycle of deep model training by computing means using a training data, a validation data, a similarity loss function, an artifact regulation loss function and a weight of loss functions to generate similarity loss and artifact regulation loss and a deep model. The method then performs a training evaluation using the similarity loss and the artifact regulation loss thus obtained to generate a training readiness output. Then, depending upon the training readiness output, the method may be terminated if certain termination criteria are met, or may perform another cycle of deep model training and training evaluation, with or without updating the weight, until the termination criteria are met. Alternatively, the deep model training in the method may be a deep adversarial model training or a bi-directional deep adversarial training.

Abstract: A computerized robust deep image transformation method performs a deep image transformation learning on multi-variation training images and corresponding desired outcome images to generate a deep image transformation model, which is applied to transform an input image to an image of higher quality mimicking a desired outcome image. A computerized robust training method for deep image integration performs a deep image integration learning on multi-modality training images and corresponding desired integrated images to generate a deep image integration model, which is applied to transform multi-modality images into a high quality integrated image mimicking a desired integrated image.

Abstract: A computerized robust deep image transformation method performs a deep image transformation learning on multi-variation training images and corresponding desired outcome images to generate a deep image transformation model, which is applied to transform an input image to an image of higher quality mimicking a desired outcome image. A computerized robust training method for deep image prediction performs a deep image prediction learning on universal modality training images and corresponding desired modality prediction images to generate a deep image prediction model, which is applied to transform universal modality images into a high quality image mimicking a desired modality prediction image.

Abstract: In a database, product image data is accumulated. A search portion acquires product image data having the image characteristics information that is the same as or similar to the image characteristics information that indicates the characteristics of the image of input image data from the database for the input image data. A search server outputs information on another product that is different from the product corresponding to the product image data together with the product image data acquired by the search portion.

Abstract: A computerized method of deep model matching for image transformation includes inputting pilot data and pre-trained deep model library into computer memories; performing a model matching scoring using the pilot data and the pre-trained deep model library to generate model matching score; and performing a model matching decision using the model matching score to generate a model matching decision output. Additional pilot data may be used to perform the model matching scoring and the model matching decision iteratively to obtain improved model matching decision output. Alternatively, the pre-trained deep model library may be pre-trained deep adversarial model library in the method.

Abstract: A clamp device is provided with a detection section which detects the pivotal position of a clamp arm. The detection section has: a cam section which is provided on a pivot shaft, includes a predetermined cam surface, and is formed so that the radial distance from the center of the pivot shaft to the cam surface varies circumferentially; and a proximity sensor which detects the position of the cam surface which varies as the pivot shaft pivots.

Abstract: A clamp device is equipped with a clamp body, a piston rod, a link mechanism, a detection mechanism for detecting a state of rotary motion of a clamp arm, and a determination unit for determining whether a state of clamping exists. The detection mechanism includes a first proximity sensor, and a knuckle joint having a first sloped surface. The determination unit, on the basis of a comparison of an output signal from the first proximity sensor to a clamp threshold value, determines whether a state of clamping exists, and on the basis of a comparison to a clamping force generation threshold value, determines whether a state of generation of clamping force exists.

Abstract: Four computerized machine learning methods for deep semantic segmentation are fast machine learning method, active machine learning method, optimal machine learning method, and optimal transfer learning method. The fast machine learning method performs a fast deep semantic segmentation learning on training images to generate a deep model. The active machine learning method performs a fast deep semantic segmentation learning on initial training images to generate a first deep model and then an active deep semantic segmentation learning to generate a second deep model. The optimal machine learning method performs a fast deep semantic segmentation learning on initial training images to generate a first deep model and then an optimal deep semantic segmentation learning to generate a second deep model. The optimal transfer learning method applies a pre-trained first deep model on transfer training images and then an optimal deep semantic segmentation transfer learning to generate a second deep model.

Abstract: Four computerized machine learning methods for deep semantic segmentation are fast machine learning method, active machine learning method, optimal machine learning method, and optimal transfer learning method. The fast machine learning method performs a fast deep semantic segmentation learning on training images to generate a deep model. The active machine learning method performs a fast deep semantic segmentation learning on initial training images to generate a first deep model and then an active deep semantic segmentation learning to generate a second deep model. The optimal machine learning method performs a fast deep semantic segmentation learning on initial training images to generate a first deep model and then an optimal deep semantic segmentation learning to generate a second deep model. The optimal transfer learning method applies a pre-trained first deep model on transfer training images and then an optimal deep semantic segmentation transfer learning to generate a second deep model.

Abstract: A clamp apparatus including two pairs of first and second clamp arms, which are supported rotatably with respect to a body and are disposed mutually in parallel. First and second cam members including respective cam surfaces are provided on ends of the first and second clamp arms. The first cam members are pressed by rollers upon lowering of a block body under a driving action of a first cylinder that makes up a drive unit. The first clamp arms are rotated to assume a clamped state. The second cam members are pressed by rollers upon lowering of a block body under a driving action of a second cylinder of the drive unit, whereby the second clamp arms are rotated to assume a clamped state.

Abstract: A check valve that can prevent sticking of a valve body and reliably perform an opening and closing action in a check vale having a simple structure and that does not use biasing means on the valve body. A check valve has support pieces of a valve body protruding in the valve body direction on a circular base seat forming an outflow window in the center inserted in a circular groove where a support member (retainer) is formed between an outflow opening and a valve seat and on an open edge of the outflow window formed in the center of the base seat, and the support pieces of the support member (retainer) that support the valve body have linear or planar support parts on the top portion for leveling at least in a radial direction of the base seat.

Abstract: A computerized method of automated hyper-parameterization for image-based deep model learning performs a deep model setup learning using initial learning images, initial truth data and a hyper-parameter setup recipe to generate deep model setup parameters, then performs a deep model learning using learning images, truth data and the generated deep model setup parameters to generate a deep model. Alternatively, the deep model learning may be a guided deep model learning. The deep model setup learning performs a deep model application, a deep quantifier calculation, and a salient hyper-parameter prediction. The hyper-parameter setup recipe may be generated by performing (a) a deep hyper-parameter mapping using application-specific learning images and application-specific truth data, (b) a salient hyper-parameter extraction, (c) a deep quantifier generation, and (d) a salient hyper-parameter prediction learning.

Abstract: A clamp apparatus is equipped with first and second clamp arms supported rotatably with respect to a body, and a drive unit having a pair of first and second pistons displaced under the supply of a pressure fluid. A driving force of the drive unit is transmitted to the first and second clamp arms through knuckle joints, which are connected to first and second piston rods, power-boost levers, and link arms. The power-boost levers are formed such that the length from a support pin toward the knuckle joint is longer than the length from the support pin toward the link arm.

Abstract: Four computerized machine learning methods for deep semantic segmentation are fast machine learning method, active machine learning method, optimal machine learning method, and optimal transfer learning method. The fast machine learning method performs a fast deep semantic segmentation learning on training images to generate a deep model. The active machine learning method performs a fast deep semantic segmentation learning on initial training images to generate a first deep model and then an active deep semantic segmentation learning to generate a second deep model. The optimal machine learning method performs a fast deep semantic segmentation learning on initial training images to generate a first deep model and then an optimal deep semantic segmentation learning to generate a second deep model. The optimal transfer learning method applies a pre-trained first deep model on transfer training images and then an optimal deep semantic segmentation transfer learning to generate a second deep model.

Abstract: A computerized robust deep image transformation method performs a deep image transformation learning on multi-variation training images and corresponding desired outcome images to generate a deep image transformation model, which is applied to transform an input image to an image of higher quality mimicking a desired outcome image. A computerized robust training method for deep image integration performs a deep image integration learning on multi-modality training images and corresponding desired integrated images to generate a deep image integration model, which is applied to transform multi-modality images into a high quality integrated image mimicking a desired integrated image.

Abstract: Provided is a production method for refining iron oxide (hematite), which has such a low sulfur content as to be used as a raw material for ironmaking from a leach residue containing iron oxide, the leach residue being produced by a high pressure acid leach (HPAL) process and being a raw material that can be cheaply and stably procured. In the method of producing (high purity) hematite for ironmaking by a process of adding an oxidant and sulfuric acid to nickel oxide ore and then leaching nickel, a leach residue obtained after the leaching of nickel is heated to 600° C. or more, and preferably 800° C. or more and 1400° C. or less.

Abstract: A clamp apparatus including first and second clamp arms, which are supported rotatably on a body, with a cover being installed to cover outer sides of the body. The cover includes openings through which portions of the first and second clamp arms are inserted and rotatably operated. An unloading tray is arranged in the body at a position below the openings in the direction of gravity. Spatter, which invades into the clamp apparatus from the openings when a welding operation is carried out on a workpiece that is clamped by the first and second clamp arms, is deposited on the unloading tray, and the spatter is eliminated by taking out the unloading tray to the exterior of the clamp apparatus.

Abstract: A check valve that can prevent sticking of a valve body and reliably perform an opening and closing action in a check vale having a simple structure and that does not use biasing means on the valve body. A check valve has support pieces of a valve body protruding in the valve body direction on a circular base seat forming an outflow window in the center inserted in a circular groove where a support member (retainer) is formed between an outflow opening and a valve seat and on an open edge of the outflow window formed in the center of the base seat, and the support pieces of the support member (retainer) that support the valve body have linear or planar support parts on the top portion for leveling at least in a radial direction of the base seat.

Abstract: A loop antenna 1 includes: a first electrode terminal 2c; a second electrode terminal 2d arranged to make a pair with the first electrode terminal 2c; and a loop-shaped member 2 which has one end connected to the first electrode terminal 2c and the other end connected to the second electrode terminal 2d, is wound a plurality of times, and is made of a conductive material. The first electrode terminal 2c and the second electrode terminal 2d are arranged so as to make a pair with respect to a center line 3 of the loop-shaped member 2. Further, the loop-shaped member 2 includes a first loop-shaped member 2a, a second loop-shaped member 2b, and an intersection part 2e. The intersection part 2e is arranged on the center line 3 in a plan view, and the loop-shaped member 2 is continuously connected and formed to be symmetrical with respect to the center line 3.