Abstract: An electronic apparatus and method for visual speech recognition based on connectionist temporal classification (CTC) loss is disclosed. The electronic apparatus receives a video that includes human speakers and generates a prediction corresponding to lip movements of the human speakers. The prediction is generated based on application of a Deep Neural Network (DNN) on the video and the DNN is trained using a CTC loss function. The electronic apparatus detects, based on the prediction, word boundaries in a sequence of characters that correspond to the lip movements and divides the video into a sequence of video clips based on the detection. Each video clip corresponds to a word spoken by the human speakers. The electronic apparatus generates a sequence of word predictions by processing the sequence of video clips and generates a sentence, or a phrase based on the generated sequence of word predictions.

Abstract: An electronic device and a method for implementation for visual speech recognition based on lip movements. The electronic device receives a set of images including one or more human speaker and applies a first machine learning (ML) model on the received set of images. The electronic device determines a first set of words spoken by the one or more human speakers based on the application of the first ML model. The determined first set of words corresponds to lip movements of the one or more human speakers. The electronic device applies a first generative Artificial Intelligence (AI) model on the determined first set of words. The electronic device predicts a first sentence corresponding to the determined first set of words spoken by the one or more human speakers, based on the application of the first generative AI model.

Abstract: An electrode sheet defect inspection device includes a vision inspection part configured to inspect defects in each tab part during a traveling process of an electrode sheet on which a plurality of tab parts is formed; a marking part configured to print a defective mark on a selected tab part in response to the tab part being determined to be defective, based on inspection information of the vision inspection part; a tab guide part configured to pass the tab parts when the electrode sheet passes through the marking part and a mark inspection part configured to inspect the defective mark printed on the tab part exposed to an outside through the opening of the tab guide part. The tab guide part has an opening for exposing a partial region of the tab parts traveling to the outside, and the tab guide part includes a transparent region adjacent to the opening.

Abstract: The present disclosure relates to a method of performing machine learning-based observation level measurement including: a log preprocessing step; a log file linkage step of processing a log file to store the log file in a HDFS, and linking the processed log file to a big data storage; a feature value extraction step of requesting an inquiry of a raw log, and extracting a feature value for a normal behavior from the inquired raw log; a model training step of normalizing the extracted feature value to level a baseline value for the normal behavior, and training a machine learning model based on the leveled baseline value; and a risk calculation step of determining, when a log that violates the leveled baseline value is detected, that an abnormal behavior is detected so as to calculate a risk for the detected abnormal behavior.

Abstract: The present invention relates to a method for attaching micro LEDs at preset mounting positions on a target substrate, comprising the steps of: (a1) disposing, at the bottom of a filled fluid, a target substrate assembly to which micro LEDs are to be attached, and preparing a water tank in which the plurality of micro LEDs float in the fluid or on the surface of the fluid on the target substrate assembly, (b1) lowering the level of the fluid in the water tank so as to lower the floating position of the plurality of micro LEDs, thereby guiding some of the plurality of micro LEDs into the plurality of apertures of the mask; and (c1) attaching, to the target substrate, some of the micro LEDs guided into the apertures.

Abstract: An exemplary embodiment of the present invention provides an apparatus for producing fermented soybean meal, which produces fermented soybean meal for monogastric animals and ruminant selectively or together.

Abstract: A refrigerator includes: a door including a door body and a door panel coupled to the door body, and a buffer member disposed at a rear surface of the door panel and provided between the door body and the door panel. The door body defines a first part at a front surface, an adhesive portion is provided at the first part between the door body and the door panel to thereby couple a rear surface of the door panel to the front surface of the door body, and the buffer member contacts a first portion of the adhesive portion and the rear surface of the door panel contacts a second portion of the adhesive portion.

Abstract: Disclosed herein is an attachment-type auxiliary device using a magnet. The attachment-type auxiliary device is used in conjunction with a product having a built-in magnetic substance. The attachment-type auxiliary device includes: a detachable part including a detachable body, and a magnet configured to be magnetically coupled with the magnetic substance; an accessory part coupled to the rear side of the detachable body, and configured to form an accommodation space and to support the product so that the product is placed in a tilted state; and a coating part made of a material having slip resistance, and formed on the front surface of the detachable body.

Abstract: The present technology relates to a battery cell jig including film-type pressure sensors, and a method of measuring a swelling of a battery cell using the same. According to the present invention, by using film-type pressure sensors, the volume of the measuring apparatus can be reduced, and abnormal degeneration of the battery cell by the pressure difference for each location of the battery cell can be prevented.

Abstract: A device risk-based trusted device verification and remote access processing includes: an inspection process start unit for starting a risk inspection process based on risk inspection information set in a device management server; a risk score calculation unit for performing a vulnerability check process and an additional check process corresponding to inspection items based on the risk inspection information to calculate a risk score for the user device through the check result; a trusted device confirmation unit, when a function of the risk score calculation unit is completely performed and the risk score is completely calculated, for transmitting the calculated risk score to the device management server, thereby allowing the device management server to determine whether the user device is a trusted device; and an access monitoring unit for monitoring data transmission and reception in real time between the user device and the business server upon remote access.

Abstract: Disclosed is a blankmask for EUV lithography, including a reflective film, a capping film and a phase shift film which are sequentially formed on a substrate. The phase shift film includes a first layer containing niobium (Nb) and chrome (Cr), and a second layer containing tantalum (Ta) and silicon (Si). In the first layer, the content of niobium (Nb) ranges from 20 to 50 at %, and the content of chrome (Cr) content ranges from 10 to 40 at %. The blankmask can implement an excellent resolution and NILS, and implement a low DtC.

Abstract: A multilayer capacitor includes a capacitor body including dielectric layers and internal electrodes alternately disposed with the dielectric layers interposed therebetween; and an external electrode disposed on the capacitor body to be connected to one or more of the internal electrodes. Porosity of ends of the internal electrodes is less than 50% on an interfacial surface between a margin of the capacitor body in a width direction the capacitor body and the internal electrodes.

Abstract: An electronic device and method for reverse image search is provided. The electronic device receives an image. The electronic device extracts, by a DNN model, a first set of image features associated with the image and generates a first feature vector based on the first set of image features. The electronic device extracts, by an image-feature detection model, a second set of image features associated with the image and generates a second feature vector based on the second set of image features. The electronic device generates a third feature vector based on combination of the first and second feature vectors. The electronic device determines a similarity metric between the third feature vector and a fourth feature vector of each of a set of pre-stored images and identifies a pre-stored image based on the similarity metric. The electronic device controls a display device to display information associated with the pre-stored image.

Abstract: A system for manufacturing edible food products according to the present disclosure includes a first food ingredient supply apparatus configured to separate a single sheet of a first food ingredient from a first food ingredient stack including a plurality of stacked first food ingredients and supply the single sheet of first food ingredient, a second food ingredient supply apparatus configured to separate a single sheet of a second food ingredient from a second food ingredient stack including a plurality of stacked second food ingredients and supply the single sheet of second food ingredient, and a pressing device configured to form an edible food product by pressing a semi-finished product formed by seating the supplied single sheet of the first food ingredient on the supplied single sheet of the second food ingredient.

Abstract: An apparatus for supplying food ingredients according to the present disclosure includes a food ingredient lifting part configured to separate and move upward a food ingredient stack from a food ingredient cassette on which food ingredients including a plurality of stacked food ingredients are seated, a food ingredient separation part configured to suck and move upward a single sheet of a food ingredient from the food ingredient stack moved upward by the food ingredient lifting part, and a horizontal movement part configured to transfer forward the single sheet of the food ingredient moved upward by the food ingredient separation part.

Abstract: A blankmask for EUV lithography includes a substrate, a reflective layer, a capping layer, and a phase shift layer. The phase shift layer is made of a material containing ruthenium (Ru) and chromium (Cr), and a total content of ruthenium (Ru) and chromium (Cr) is 50 to 100 at %. The phase shift layer may further contain boron (B) or nitrogen (N). The phase shift layer of the present invention has a high relative reflectance (relative reflectance with respect to a reflectance of the reflective layer under the phase shift layer) with respect to a tantalum (Ta)-based phase shift layer and has a phase shift amount of 170 to 230°. It is possible to obtain excellent resolution when finally manufacturing a pattern of 7 nm or less by using a photomask manufactured using such a blankmask.

Abstract: A cutting apparatus according to the present disclosure includes a cutting roller including a cutting body having a cylindrical shape and configured to rotate about an axis defined in a leftward/rightward direction, and cutting blades protruding outward in a radial direction of the cutting body further than a surface of the cutting body to cut an edible food product provided in a forward/rearward direction, and a cutting base part disposed at a position facing the cutting roller based on the edible food product to support the edible food product to be cut by the cutting roller.