Abstract: One embodiment of the present invention sets forth a technique for performing a labeling task. The technique includes generating a multi-scale representation of an image as input to a machine learning model. The technique also includes performing one or more operations that apply the machine learning model to the multi-scale representation of the image to produce a semantic segmentation comprising predictions of labels for regions of pixels in the image. The technique further includes outputting, in a user interface, the semantic segmentation for use in assisting a user in specifying the labels for the pixels in the image.

Abstract: A detection system includes a photodetector, a spatial light modulator (SLM) device optically coupled to the photodetector, and a controller. The controller is operatively connected to the SLM device and is disposed in communication with the photodetector and a memory. The memory has instructions recorded on the memory that cause the controller to communicate a SLM pattern sequence to the SLM device, modulate illumination incident on the SLM device according to the SLM pattern sequence to generate an illumination pulse sequence, and receive an intensity-time profile from the photodetector corresponding to the SLM pattern sequence. The instructions also cause the controller to signal, with spatial specificity, presence of flame or gas when the intensity-time profile indicates that flame or gas is present within a field of view of the detection system. Detection methods and computer program products are also described.

Abstract: Methods and systems for gas detection are provided. Aspects include determining a target gas for detection in a sampling chamber, determining one or more target characteristics of light based on the target gas, operating a light source to transmit the light through a sampling chamber to an optical element, operating an active optical element to modulate the light based on the one or more target characteristics of the light, operating a filter to receive the light from the optical element and separate the light in to a first light portion and a second light portion, operating the photodetector to receive the first light portion from the filter, and analyzing the first light portion to determine a presence of the target gas in the sampling chamber.