The key challenge of multi-view indoor 3D object detection is to infer accurate geometry information from images for precise 3D detection. Previous method relies on NeRF for geometry reasoning.

Researchers from the Institute of Electrical and Electronics Engineers (IEEE) have developed a method to increase the authenticity of low-cost, projection-based augmented reality installations, through special glasses that cause projected 3D images to go in and out of focus in the same way that they would if the objects were real, overcoming a critical perceptual hurdle for practical usage of projection systems in controlled environments. The IEEE system recreates depth planes for projected real and CGI imagery that will be superimposed into rooms. In this case, three CGI Stanford bunnies are being superimposed at the same depth plane as three real world objects, and their blurriness is controlled by where the viewer is looking and focusing. The system uses electrically focus-tunable lenses (ETL) embedded into the viewer's glasses (which are in any case necessary to separate the two image streams into a convincing, integrated 3D experience), and which communicate with the projection system, which then automatically changes the level of blurriness of the projected image seen by the viewer. The ETL lenses report back information about the user's focal attention and sets the level of blurriness on a per-plane basis for the rendering of the projected geometry.

In ultrasound (US) imaging, individual channel RF measurements are back-propagated and accumulated to form an image after applying specific delays. While this time reversal is usually implemented using a hardware- or software-based delay-and-sum (DAS) beamformer, the performance of DAS decreases rapidly in situations where data acquisition is not ideal. Herein, for the first time, we demonstrate that a single data-driven beamformer designed as a deep neural network can directly process sub-sampled RF data acquired at different sampling rates to generate high quality US images. In particular, the proposed deep beamformer is evaluated for two distinct acquisition schemes: focused ultrasound imaging and planewave imaging. Experimental results showed that the proposed deep beamformer exhibit significant performance gain for both focused and planar imaging schemes, in terms of contrast-to-noise ratio and structural similarity.

The classical computational model for stereo vision incorporates a uniqueness inhibition constraint to enforce a one-to-one feature match, thereby sacrificing the ability to handle transparency. Critics of the model disregard the uniqueness constraint and argue that the smoothness constraint can provide the excitation support required for transparency computation. However, this modification fails in neighborhoods with sparse features. We propose a Bayesian approach to stereo vision with priors favoring cohesive over transparent surfaces. The disparity and its segmentation into a multi-layer "depth planes" representation are simultaneously computed. The smoothness constraint propagates support within each layer, providing mutual excitation for non-neighboring transparent or partially occluded regions. Test results for various random-dot and other stereograms are presented.

The classical computational model for stereo vision incorporates a uniqueness inhibition constraint to enforce a one-to-one feature match, thereby sacrificing the ability to handle transparency. Critics of the model disregard the uniqueness constraint and argue that the smoothness constraint can provide the excitation support required for transparency computation. However, this modification fails in neighborhoods with sparse features. We propose a Bayesian approach to stereo vision with priors favoring cohesive over transparent surfaces. The disparity and its segmentation into a multi-layer "depth planes" representation are simultaneously computed. The smoothness constraint propagates support within each layer, providing mutual excitation for non-neighboring transparent or partially occluded regions. Test results for various random-dot and other stereograms are presented.

The classical computational model for stereo vision incorporates a uniqueness inhibition constraint to enforce a one-to-one feature match, thereby sacrificing the ability to handle transparency. Critics ofthe model disregard the uniqueness constraint and argue that the smoothness constraint can provide the excitation support required for transparency computation.