Abstract: Described are systems, techniques, and processes for pathology detection in radiological images. A process can include obtaining a radiological image corresponding to an imaged anatomical area. Based on processing the radiological image using a semantic segmentation neural network, a target map can be generated corresponding to a plurality of candidate anatomical defect locations within the cropped radiological image. At least one volume of interest (VOI) can be generated centered around a particular candidate anatomical defect location within the cropped radiological image. A classification neural network can be used to classify the particular candidate anatomical defect location within the cropped radiological image, wherein classifying the particular candidate anatomical defect location includes determining a pathology associated with the particular candidate anatomical defect location.

Abstract: Described are techniques for image processing. For instance, a process can include obtaining a plurality of labeled input images and determining a threshold percentage associated with the plurality of labeled input images, indicative of a percentage of correctly labeled input images. The process can include determining a respective self-influence for each respective labeled input image included in the plurality of input images and generating a respective self-influence weight for each respective labeled input image, based on the respective self-influence and the threshold percentage associated with each respective labeled input image. The process can include determining one or more loss values using a loss function associated with training a machine learning network based on using the plurality of labeled input images as a training data set, wherein the loss function determines the one or more loss values based on weighting each respective labeled input image by its respective self-influence weight.

Abstract: In an embodiment, a computer-implemented process comprises accessing a plurality of digitally stored, unstructured medical diagnostic data; digitally displaying a first subset of the medical diagnostic data, the first subset of the medical diagnostic data including at least a first set of diagnostic reports, using a computer display device, concurrently with digitally displaying one or more quality control checklists that are specific to a medical discipline represented in the first set of diagnostic reports; receiving digital input specifying one or more errors in the first set of diagnostic reports and digitally storing the digital input in association with the first subset of medical diagnostic data; training a hierarchical Bayesian machine learning model using the digital input and the first subset of medical diagnostic data; evaluating the hierarchical Bayesian machine learning model, after training, for a second subset of the medical diagnostic data, the second subset being different from the first subs

Abstract: Described are techniques for determining an accurate ground-truth Bayesian Inter-Reviewer Agreement Rate (BIRAR) from unreliable sources. For instance, a process can include obtaining an initial set of diagnostic imaging exams, wherein each diagnostic imaging exam includes a severity grade associated with an initial radiologist. For each diagnostic imaging exam, two or more secondary quality assurance (QA) reviews can be obtained for each diagnostic imaging exam, wherein the secondary QA reviews are associated with QA'ing radiologists different than the initial radiologist. One or more inter-reviewer agreement rates can be determined for the QA'ing radiologists, based on the secondary QA reviews associated with the QA'ing radiologists.

Abstract: Described are techniques for uncertainty-aware anatomical landmark detection, using, for example, a deep reinforcement learning (DRL) anatomical landmark detection agent. For instance, a process can include generating one or more image features for an input medical image using a first sub-network of the anatomical landmark detection agent. A softmax layer of a second sub-network of the anatomical landmark detection agent can generate a plurality of discrete Q-value distributions for a set of allowable actions associated with movement of the agent within the medical image. An anatomical landmark location within the medical image can be predicted using the discrete Q-value distributions. An uncertainty can be determined for the predicted anatomical landmark location, based on an average full width half maximum (FWHM) calculated for the plurality of discrete Q-value distributions.

Abstract: In an embodiment, a computer-implemented process comprises accessing a plurality of digitally stored, unstructured medical diagnostic data; digitally displaying a first subset of the medical diagnostic data, the first subset of the medical diagnostic data including at least a first set of diagnostic reports, using a computer display device, concurrently with digitally displaying one or more quality control checklists that are specific to a medical discipline represented in the first set of diagnostic reports; receiving digital input specifying one or more errors in the first set of diagnostic reports and digitally storing the digital input in association with the first subset of medical diagnostic data; training a hierarchical Bayesian machine learning model using the digital input and the first subset of medical diagnostic data; evaluating the hierarchical Bayesian machine learning model, after training, for a second subset of the medical diagnostic data, the second subset being different from the first subs

Abstract: A light assembly for a refrigerator drawer includes a light source assembly, a reflector and a lens assembly keyed one to another for positioning, and secured to each other as a preassembled unit. The preassembled light assembly is snap fit into an opening in a panel of the drawer.

Abstract: A detection apparatus for detecting refrigerator door openings is coupled to at least one switch configured to be activated by a door opening. When the door is opened, the switch is activated and inputs a signal to the detection apparatus. The detection apparatus rectifies the signal; and phase-shifts the rectified signal so that it leads or lags the line voltage. The shifted output signal is fed to a processor that detects the opening of the door based upon the shifted signal. Signals output by a plurality of switches that generate a signal when activated mixed using an opto-coupler. Relative impedance of the lead and lag circuits may be adjusted to differentiate a phase shift of one shifted signal relative to another signal. The processor converts a value in degrees of phase shifting of the mixed signal to a time value, and based upon the time value, the processor determines which of the doors is open.

Abstract: An electrical assembly for mounting an electrical device, such as a lamp socket, in an appliance has first and second components on opposite sides of a wall of the appliance. The first component is attached to the wall and closes a hole through the wall on a first side thereof. The second component is attached to the first component and closes the hole on an opposite side of the wall. Complementary quick connect electrical terminals are provided in the first and second components whereby electrical connection is established as the first and second components are attached one to the other.

Abstract: A control system for an appliance utilizing a plurality of peripheral devices includes a controller and a communications bus coupled to the plurality of peripheral devices and to the controller. The controller is configured to receive data from the peripheral devices over the communications bus and transmit control operations to the peripheral devices based on the data received. The communications bus reduces overall wiring requirements due to elimination or reduction of power wiring of the peripheral devices, facilitates distributed displays and human machine interfaces, and allows model differentiation with late point identification by a population of bus connections with a sensor, display, interface, or a blank. A data collision detection scheme is also provided to ensure proper communication along the bus.

Abstract: An adaptive defrost control includes a microcomputer for controlling the initiation and termination of a defrost operation based, at least in part, on opening of the fresh food door and the freezer door. The control monitors the compressor run time and the fresh food and freezer door open times and adjusts the time until defrost accordingly. A pre-chill operation cools the freezer prior to defrosting the evaporator coils so that the defrost heat will have less of an affect on the maximum temperature after defrost, and enables the freezer compartment to be maintained at a temperature a few degrees higher than with known refrigerators, providing energy savings.

Abstract: A method for controlling speed in a pulse-width-modulation-controlled motor powered by a load voltage source is provided. The method includes the steps of measuring the motor load voltage, and setting a pulse-width-modulation duty cycle based on the measured voltage.

Abstract: Systems and methods for controlling locking of a washing machine lid are described. The washing machine includes an agitation element and a basket. The method, in one example embodiment, comprises the steps of sensing a spin speed associated with a spin speed of at least one of the agitation element and the basket, and causing the lid to be locked when the sensed spin speed exceed a first predetermined speed.

Abstract: An electronic controller for an automatic icemaker includes a thermistor coupled to an icemaker mold, a first transducer coupled to a rake, a second transducer coupled to a feeler arm, and a processor. Time and temperature dependent calculations are summed at successive time intervals and compared to a predetermined value to determine when ice is frozen and ready for harvesting. Harvesting is delayed a pre-selected time based upon the position of the feeler arm to avoid premature harvesting. A harvest fix algorithm is executed when ice is not harvested within a predetermined time period by operating a heater at an elevated temperature and by de-energizing a motor. A water fill algorithm determines whether the mold is full of water based upon a change in the thermistor reading.

Abstract: A method for controlling speed in a pulse-width-modulation-controlled motor powered by a load voltage source is provided. The method includes the steps of measuring the motor load voltage, and setting a pulse-width-modulation duty cycle based on the measured voltage.

Abstract: A control system for an appliance utilizing a plurality of peripheral devices includes a controller and a communications bus coupled to the plurality of peripheral devices and to the controller. The controller is configured to receive data from the peripheral devices over the communications bus and transmit control operations to the peripheral devices based on the data received. The communications bus reduces overall wiring requirements due to elimination or reduction of power wiring of the peripheral devices, facilitates distributed displays and human machine interfaces, and allows model differentiation with late point identification by a population of bus connections with a sensor, display, interface, or a blank. A data collision detection scheme is also provided to ensure proper communication along the bus.

Abstract: A detection apparatus for detecting refrigerator door openings is coupled to at least one switch configured to be activated by a door opening. When the door is opened, the switch is activated and inputs a signal to the detection apparatus. The detection apparatus rectifies the signal; and phase-shifts the rectified signal so that it leads or lags the line voltage. The shifted output signal is fed to a processor that detects the opening of the door based upon the shifted signal. Signals output by a plurality of switches that generate a signal when activated mixed using an opto-coupler. Relative impedance of the lead and lag circuits may be adjusted to differentiate a phase shift of one shifted signal relative to another signal. The processor converts a value in degrees of phase shifting of the mixed signal to a time value, and based upon the time value, the processor determines which of the doors is open.

Abstract: An improved system and method for controlling the temperature of a resistance heater in a heating, ventilation and air-conditioning (“HVAC”) system is disclosed. The control circuit and method includes three separate thermostats per heater leg and are designed to trip at three, sequential pre-selected temperatures to address certain issues associated with prior art designs for thermal overload protection. One of the thermostats is of an automatically resettable type wherein the remaining two are of a “one-shot” design and will remain open until there is human intervention.

Abstract: An air flow control arrangement for a side-by-side refrigerator which results in temperature variations, both spatial and temporal, of approximately about .+-.1.degree. F. in the fresh food compartment, is described. Such air flow control also is believed to even further reduce the internal "sweat" in the fresh food compartment. In one embodiment, and with respect to a side-by-side refrigerator, the middle duct in the center mullion wall is closed off, and a duct is added in the center mullion wall, or could be installed as a separate duct outside the mullion, connecting a supply port in the top, back, left hand corner in fresh food compartment to a position in the evaporator chamber. The added duct also seals off direct airflow between the evaporator chamber and the fresh food compartment.