A non-transitory computer readable medium includes instructions that, when executed by at least one processor, cause the at least one processor to perform artificial-intelligence-based technical support operations. The operations may include receiving over at least one network first audio signals including speech data associated with a technical support session and first image signals including image data associated with a product for which support is sought from a mobile communications device, analyzing the first audio signals and the first image signals using artificial intelligence, aggregating the analysis thereof, accessing at least one data structure to identify an image capture instruction, presenting the image capture instruction including a direction to alter and capture second image signals of a structure identified in the first image signals to the mobile communications device, receiving from the mobile communications device second image signals, analyzing the same using artificial intelligence, and determining a technical support resolution status.

A method and system for detecting a collision by an autonomous robot based on a multi-sensor long short-term memory (LSTM) are disclosed. The method includes generating an input of an LSTM model by combining an input image received from an autonomous robot, light detection and ranging (LiDAR) distance information, and acceleration information, learning a collision alert situation by inputting the input to the LSTM model, and determining a collision situation using an output of the LSTM model and a fully connected neural network (FNN) model.

Discussed herein are systems, devices, and methods for automatic target recognition based on a non-visible input image. A method can include providing, as input to a first machine learning (ML) model for object classification, pixel data of a non-visible image, the first ML model including an encoder from a second ML model, the second ML model trained to generate a visible image representation of an input non-visible image, and receiving, from the first ML model, data indicating one or more objects present in the non-visible image.

Methods, techniques, apparatus, and algorithms are described for robustly measuring real-world distances using any mobile device equipped with an accelerometer and monocular camera. A general software implementation processes 2D video, precisely tracking points of interest across frames to estimate the unsealed trajectory of the device, which is used to correct the device's inertially derived trajectory. The visual and inertial trajectories are then aligned in scale space to estimate the physical distance travelled by the device and the true distance between the visually tracked points.

An information processing apparatus capable of detecting a plane constituting a movement-enabling region. A normal direction of a plane constituting a road surface is detected on the basis of polarized images in multiple polarizing directions acquired by a polarization camera. A laser ranging sensor measures a distance to a point on the road surface so as to measure a position of the point. The plane constituting the road surface is identified on the basis of information regarding the normal direction of the plane constituting the road surface and information regarding the position of the point on the road surface.

An image processing apparatus includes a change detection unit configured to detect phase changes in multiple predetermined directions from among phase changes in a luminance image in units of mutually different resolutions, and a reliability estimation unit configured to estimate reliability of the detected phase change based on temporal amplitude change information in the multiple directions determined in the luminance image. The reliability estimation unit may estimate the reliability using an amplitude change of multiple resolutions and using a value of an amplitude change equal to or greater than a predetermined threshold value among images having multiple resolutions. The reliability may become a greater value as the amplitude change becomes larger.

Provided are an image recognition device and an image recognition method capable of improving subject recognition accuracy. The image recognition device (image sensor 1) according to the present disclosure includes an imaging unit (10) and a recognition unit (14). The imaging unit (10) generates image data by capturing a plurality of images in different wavelength bands using imaging pixels (R, G, B, IR) receiving light in four or more types of wavelength bands. The recognition unit (14) recognizes a subject from each of the plurality of pieces of image data for each of the wavelength bands.

A skeleton recognition method for a computer to execute a process includes acquiring distance images from each of a plurality of sensors that sense a subject from a plurality of directions; acquiring joint information that includes joint positions of the subject for each of the plurality of sensors by using a machine learning model that estimates the joint positions from the distance images; generating skeleton information that represents three-dimensional coordinates by integrating the joint information; and outputting the skeleton information of the subject.

A data extraction unit (23) extracts, out of image data obtained by photographing a photographing region with a photographing device (41), image data of a region including a detection target region, as target data, and extracts, out of the target data, image data of an enlarging region, as partial data. A size modification unit (24) size-modifies each of the target data and the partial data to a request size requested by an object detection model being a model that detects an object from image data. An object detection unit (25) inputs each of the size-modified target data and the size-modified partial data to the object detection model, and detects a target object from each of the target data and the partial data.

Provided is a multispectral imaging device for providing precise tracking and verification. The imaging device may configure a first filter for a sensor, and may determine first spectral properties of a target object based on a first image of the target object generated from visible light passing through the first filter onto the sensor. The imaging device may configure a different second filter for the sensor, and may determine second spectral properties of the target object based on a second image of the target object generated from the non-visible light passing through the second filter onto the sensor. The imaging device may align the second spectral properties of the second image with the first spectral properties of the first image, and may present the first spectral properties with the second spectral properties in a single composite image of the target object.